feat: add DXT packaging support (experimental)

- Add comprehensive manifest.json with all 19 Basic Memory MCP tools
- Implement automated DXT build process via justfile (just dxt)
- Add DXT assets (icons and screenshots)
- Create .dxtignore for excluding development files
- Update .gitignore to exclude generated DXT bundle files

Note: DXT implementation blocked by Anthropic DXT v0.1 limitations:
- Binary permissions issues (anthropics/dxt#12)
- Platform-specific dependency problems (anthropics/dxt#17)
- Native module loading errors with pydantic_core

See issue #193 for detailed analysis and future implementation plan.

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>

.claude/commands/release/release.md

@@ -15,16 +15,10 @@ Create a stable release using the automated justfile target with comprehensive v
 You are an expert release manager for the Basic Memory project. When the user runs `/release`, execute the following steps:
 
 ### Step 1: Pre-flight Validation
-
-#### Version Check
-1. Check current version in `src/basic_memory/__init__.py`
-2. Verify new version format matches `v\d+\.\d+\.\d+` pattern
-3. Confirm version is higher than current version
-
-#### Git Status
-1. Check current git status for uncommitted changes
-2. Verify we're on the `main` branch
-3. Confirm no existing tag with this version
+1. Verify version format matches `v\d+\.\d+\.\d+` pattern
+2. Check current git status for uncommitted changes  
+3. Verify we're on the `main` branch
+4. Confirm no existing tag with this version
 
 #### Documentation Validation
 1. **Changelog Check**
@@ -45,111 +39,19 @@ The justfile target handles:
 - ✅ Version update in `src/basic_memory/__init__.py`
 - ✅ Automatic commit with proper message
 - ✅ Tag creation and pushing to GitHub
-- ✅ Release workflow trigger (automatic on tag push)
-
-The GitHub Actions workflow (`.github/workflows/release.yml`) then:
-- ✅ Builds the package using `uv build`
-- ✅ Creates GitHub release with auto-generated notes
-- ✅ Publishes to PyPI
-- ✅ Updates Homebrew formula (stable releases only)
+- ✅ Release workflow trigger
 
 ### Step 3: Monitor Release Process
-1. Verify tag push triggered the workflow (should start automatically within seconds)
-2. Monitor workflow progress at: https://github.com/basicmachines-co/basic-memory/actions
-3. Watch for successful completion of both jobs:
-   - `release` - Builds package and publishes to PyPI
-   - `homebrew` - Updates Homebrew formula (stable releases only)
-4. Check for any workflow failures and investigate logs if needed
+1. Check that GitHub Actions workflow starts successfully
+2. Monitor workflow completion at: https://github.com/basicmachines-co/basic-memory/actions
+3. Verify PyPI publication
+4. Test installation: `uv tool install basic-memory`
 
 ### Step 4: Post-Release Validation
-
-#### GitHub Release
-1. Verify GitHub release is created at: https://github.com/basicmachines-co/basic-memory/releases/tag/<version>
-2. Check that release notes are auto-generated from commits
-3. Validate release assets (`.whl` and `.tar.gz` files are attached)
-
-#### PyPI Publication
-1. Verify package published at: https://pypi.org/project/basic-memory/<version>/
-2. Test installation: `uv tool install basic-memory`
-3. Verify installed version: `basic-memory --version`
-
-#### Homebrew Formula (Stable Releases Only)
-1. Check formula update at: https://github.com/basicmachines-co/homebrew-basic-memory
-2. Verify formula version matches release
-3. Test Homebrew installation: `brew install basicmachines-co/basic-memory/basic-memory`
-
-#### MCP Registry Publication
-
-After PyPI release is published, update the MCP registry:
-
-1. **Verify PyPI Release**
-   - Confirm package is live: https://pypi.org/project/basic-memory/<version>/
-   - The `server.json` version was auto-updated by `just release`
-
-2. **Publish to MCP Registry**
-   ```bash
-   cd /Users/drew/code/basic-memory
-   mcp-publisher publish
-   ```
-
-   If not authenticated:
-   ```bash
-   mcp-publisher login github
-   # Follow device authentication flow
-   mcp-publisher publish
-   ```
-
-3. **Verify Publication**
-   ```bash
-   curl "https://registry.modelcontextprotocol.io/v0.1/servers?search=basic-memory"
-   ```
-
-**Note:** The `mcp-publisher` CLI can be installed via Homebrew (`brew install mcp-publisher`) or from GitHub releases.
-
-#### Website Updates
-
-**1. basicmachines.co** (`/Users/drew/code/basicmachines.co`)
-   - **Goal**: Update version number displayed on the homepage
-   - **Location**: Search for "Basic Memory v0." in the codebase to find version displays
-   - **What to update**:
-     - Hero section heading that shows "Basic Memory v{VERSION}"
-     - "What's New in v{VERSION}" section heading
-     - Feature highlights array (look for array of features with title/description)
-   - **Process**:
-     1. Pull latest from GitHub: `git pull origin main`
-     2. Create release branch: `git checkout -b release/v{VERSION}`
-     3. Search codebase for current version number (e.g., "v0.16.1")
-     4. Update version numbers to new release version
-     5. Update feature highlights with 3-5 key features from this release (extract from CHANGELOG.md)
-     6. Commit changes: `git commit -m "chore: update to v{VERSION}"`
-     7. Push branch: `git push origin release/v{VERSION}`
-   - **Deploy**: Follow deployment process for basicmachines.co
-
-**2. docs.basicmemory.com** (`/Users/drew/code/docs.basicmemory.com`)
-   - **Goal**: Add new release notes section to the latest-releases page
-   - **File**: `src/pages/latest-releases.mdx`
-   - **What to do**:
-     1. Pull latest from GitHub: `git pull origin main`
-     2. Create release branch: `git checkout -b release/v{VERSION}`
-     3. Read the existing file to understand the format and structure
-     4. Read `/Users/drew/code/basic-memory/CHANGELOG.md` to get release content
-     5. Add new release section **at the top** (after MDX imports, before other releases)
-     6. Follow the existing pattern:
-        - Heading: `## [v{VERSION}](github-link) — YYYY-MM-DD`
-        - Focus statement if applicable
-        - `<Info>` block with highlights (3-5 key items)
-        - Sections for Features, Bug Fixes, Breaking Changes, etc.
-        - Link to full changelog at the end
-        - Separator `---` between releases
-     7. Commit changes: `git commit -m "docs: add v{VERSION} release notes"`
-     8. Push branch: `git push origin release/v{VERSION}`
-   - **Source content**: Extract and format sections from CHANGELOG.md for this version
-   - **Deploy**: Follow deployment process for docs.basicmemory.com
-
-**4. Announce Release**
-   - Post to Discord community if significant changes
-   - Update social media if major release
-   - Notify users via appropriate channels
+1. Verify GitHub release is created automatically
+2. Check PyPI publication
+3. Validate release assets
+4. Update any post-release documentation
 
 ## Pre-conditions Check
 Before starting, verify:
@@ -172,28 +74,19 @@ Before starting, verify:
 🏷️  Tag: v0.13.2
 📋 GitHub Release: https://github.com/basicmachines-co/basic-memory/releases/tag/v0.13.2
 📦 PyPI: https://pypi.org/project/basic-memory/0.13.2/
-🍺 Homebrew: https://github.com/basicmachines-co/homebrew-basic-memory
-🔌 MCP Registry: https://registry.modelcontextprotocol.io
 🚀 GitHub Actions: Completed
 
-Install with pip/uv:
-  uv tool install basic-memory
-
-Install with Homebrew:
-  brew install basicmachines-co/basic-memory/basic-memory
+Install with:
+uv tool install basic-memory
 
 Users can now upgrade:
-  uv tool upgrade basic-memory
-  brew upgrade basic-memory
+uv tool upgrade basic-memory
 ```
 
 ## Context
 - This creates production releases used by end users
 - Must pass all quality gates before proceeding
 - Uses the automated justfile target for consistency
-- Version is automatically updated in `__init__.py` and `server.json`
+- Version is automatically updated in `__init__.py`
 - Triggers automated GitHub release with changelog
-- Package is published to PyPI for `pip` and `uv` users
-- Homebrew formula is automatically updated for stable releases
-- MCP Registry is updated manually via `mcp-publisher publish`
-- Supports multiple installation methods (uv, pip, Homebrew)
+- Leverages uv-dynamic-versioning for package version management

.claude/commands/spec.md

@@ -1,51 +0,0 @@
----
-allowed-tools: mcp__basic-memory__write_note, mcp__basic-memory__read_note, mcp__basic-memory__search_notes, mcp__basic-memory__edit_note
-argument-hint: [create|status|show|review] [spec-name]
-description: Manage specifications in our development process
----
-
-## Context
-
-Specifications are managed in the Basic Memory "specs" project. All specs live in a centralized location accessible across all repositories via MCP tools.
-
-See SPEC-1 and SPEC-2 in the "specs" project for the full specification-driven development process.
-
-Available commands:
-- `create [name]` - Create new specification
-- `status` - Show all spec statuses
-- `show [spec-name]` - Read a specific spec
-- `review [spec-name]` - Review implementation against spec
-
-## Your task
-
-Execute the spec command: `/spec $ARGUMENTS`
-
-### If command is "create":
-1. Get next SPEC number by searching existing specs in "specs" project
-2. Create new spec using template from SPEC-2
-3. Use mcp__basic-memory__write_note with project="specs"
-4. Include standard sections: Why, What, How, How to Evaluate
-
-### If command is "status":
-1. Use mcp__basic-memory__search_notes with project="specs"
-2. Display table with spec number, title, and progress
-3. Show completion status from checkboxes in content
-
-### If command is "show":
-1. Use mcp__basic-memory__read_note with project="specs"
-2. Display the full spec content
-
-### If command is "review":
-1. Read the specified spec and its "How to Evaluate" section
-2. Review current implementation against success criteria with careful evaluation of:
-   - **Functional completeness** - All specified features working
-   - **Test coverage analysis** - Actual test files and coverage percentage
-     - Count existing test files vs required components/APIs/composables
-     - Verify unit tests, integration tests, and end-to-end tests
-     - Check for missing test categories (component, API, workflow)
-   - **Code quality metrics** - TypeScript compilation, linting, performance
-   - **Architecture compliance** - Component isolation, state management patterns
-   - **Documentation completeness** - Implementation matches specification
-3. Provide honest, accurate assessment - do not overstate completeness
-4. Document findings and update spec with review results using mcp__basic-memory__edit_note
-5. If gaps found, clearly identify what still needs to be implemented/tested

.claude/commands/test-live.md

@@ -11,7 +11,7 @@ All test results are recorded as notes in a dedicated test project.
 **Parameters:**
 - `phase` (optional): Specific test phase to run (`recent`, `core`, `features`, `edge`, `workflows`, `stress`, or `all`)
 - `recent` - Focus on recent changes and new features (recommended for regular testing)
-- `core` - Essential tools only (Tier 1: write_note, read_note, search_notes, edit_note, list_memory_projects, recent_activity)
+- `core` - Essential tools only (Tier 1: write_note, read_note, search_notes, edit_note, list_projects, switch_project)
 - `features` - Core + important workflows (Tier 1 + Tier 2)
 - `all` - Comprehensive testing of all tools and scenarios
 
@@ -24,66 +24,30 @@ When the user runs `/project:test-live`, execute comprehensive test plan:
 
 ### **Tier 1: Critical Core (Always Test)**
 1. **write_note** - Foundation of all knowledge creation
-2. **read_note** - Primary knowledge retrieval mechanism
+2. **read_note** - Primary knowledge retrieval mechanism  
 3. **search_notes** - Essential for finding information
 4. **edit_note** - Core content modification capability
-5. **list_memory_projects** - Project discovery and session guidance
-6. **recent_activity** - Project discovery mode and activity analysis
+5. **list_memory_projects** - Project discovery and status
+6. **switch_project** - Context switching for multi-project workflows
 
 ### **Tier 2: Important Workflows (Usually Test)**
-7. **build_context** - Conversation continuity via memory:// URLs
-8. **create_memory_project** - Essential for project setup
-9. **move_note** - Knowledge organization
-10. **sync_status** - Understanding system state
-11. **delete_project** - Project lifecycle management
+7. **recent_activity** - Understanding what's changed
+8. **build_context** - Conversation continuity via memory:// URLs
+9. **create_memory_project** - Essential for project setup
+10. **move_note** - Knowledge organization
+11. **sync_status** - Understanding system state
 
 ### **Tier 3: Enhanced Functionality (Sometimes Test)**
 12. **view_note** - Claude Desktop artifact display
 13. **read_content** - Raw content access
 14. **delete_note** - Content removal
 15. **list_directory** - File system exploration
-16. **edit_note** (advanced modes) - Complex find/replace operations
+16. **set_default_project** - Configuration
+17. **delete_project** - Administrative cleanup
 
 ### **Tier 4: Specialized (Rarely Test)**
-17. **canvas** - Obsidian visualization (specialized use case)
-18. **MCP Prompts** - Enhanced UX tools (ai_assistant_guide, continue_conversation)
-
-## Stateless Architecture Testing
-
-### **Project Discovery Workflow (CRITICAL)**
-Test the new stateless project selection flow:
-
-1. **Initial Discovery**
-   - Call `list_memory_projects()` without knowing which project to use
-   - Verify clear session guidance appears: "Next: Ask which project to use"
-   - Confirm removal of CLI-specific references
-
-2. **Activity-Based Discovery**
-   - Call `recent_activity()` without project parameter (discovery mode)
-   - Verify intelligent project suggestions based on activity
-   - Test guidance: "Should I use [most-active-project] for this task?"
-
-3. **Session Tracking Validation**
-   - Verify all tool responses include `[Session: Using project 'name']`
-   - Confirm guidance reminds about session-wide project tracking
-
-4. **Single Project Constraint Mode**
-   - Test MCP server with `--project` parameter
-   - Verify all operations constrained to specified project
-   - Test project override behavior in constrained mode
-
-### **Explicit Project Parameters (CRITICAL)**
-All tools must require explicit project parameters:
-
-1. **Parameter Validation**
-   - Test all Tier 1 tools require `project` parameter
-   - Verify clear error messages for missing project
-   - Test invalid project name handling
-
-2. **No Session State Dependencies**
-   - Confirm no tool relies on "current project" concept
-   - Test rapid project switching within conversation
-   - Verify each call is truly independent
+18. **canvas** - Obsidian visualization (specialized use case)
+19. **MCP Prompts** - Enhanced UX tools (ai_assistant_guide, continue_conversation)
 
 ### Pre-Test Setup
 
@@ -108,7 +72,7 @@ Run the bash `date` command to get the current date/time.
    Purpose: Record all test observations and results
    ```
 
-Make sure to use the newly created project for all subsequent test operations by specifying it in the `project` parameter of each tool call.
+Make sure to switch to the newly created project with the `switch_project()` tool.
 
 4. **Baseline Documentation**
    Create initial test session note with:
@@ -179,42 +143,46 @@ Test essential MCP tools that form the foundation of Basic Memory:
 - ⚠️ Error scenarios (invalid operations)
 
 **5. list_memory_projects Tests (Critical):**
-- ✅ Display all projects with clear session guidance
-- ✅ Project discovery workflow prompts
-- ✅ Removal of CLI-specific references
+- ✅ Display all projects with status indicators
+- ✅ Current and default project identification
 - ✅ Empty project list handling
-- ✅ Single project constraint mode display
+- ✅ Project metadata accuracy
 
-**6. recent_activity Tests (Critical - Discovery Mode):**
-- ✅ Discovery mode without project parameter
-- ✅ Intelligent project suggestions based on activity
-- ✅ Guidance prompts for project selection
-- ✅ Session tracking reminders in responses
-- ⚠️ Performance with multiple projects
+**6. switch_project Tests (Critical):**
+- ✅ Switch between existing projects
+- ✅ Context preservation during switch
+- ⚠️ Invalid project name handling
+- ✅ Confirmation of successful switch
 
 ### Phase 2: Important Workflows (Tier 2 Tools)
 
-**7. build_context Tests (Important):**
+**7. recent_activity Tests (Important):**
+- ✅ Various timeframes ("today", "1 week", "1d")
+- ✅ Type filtering capabilities
+- ✅ Empty project scenarios
+- ⚠️ Performance with many recent changes
+
+**8. build_context Tests (Important):**
 - ✅ Different depth levels (1, 2, 3+)
 - ✅ Various timeframes for context
 - ✅ memory:// URL navigation
 - ⚠️ Performance with complex relation graphs
 
-**8. create_memory_project Tests (Important):**
+**9. create_memory_project Tests (Important):**
 - ✅ Create projects dynamically
 - ✅ Set default during creation
 - ✅ Path validation and creation
 - ⚠️ Invalid paths and names
 - ✅ Integration with existing projects
 
-**9. move_note Tests (Important):**
+**10. move_note Tests (Important):**
 - ✅ Move within same project
 - ✅ Cross-project moves with detection (#161)
 - ✅ Automatic folder creation
 - ✅ Database consistency validation
 - ⚠️ Special characters in paths
 
-**10. sync_status Tests (Important):**
+**11. sync_status Tests (Important):**
 - ✅ Background operation monitoring
 - ✅ File synchronization status
 - ✅ Project sync state reporting
@@ -222,31 +190,36 @@ Test essential MCP tools that form the foundation of Basic Memory:
 
 ### Phase 3: Enhanced Functionality (Tier 3 Tools)
 
-**11. view_note Tests (Enhanced):**
+**12. view_note Tests (Enhanced):**
 - ✅ Claude Desktop artifact display
 - ✅ Title extraction from frontmatter
 - ✅ Unicode and emoji content rendering
 - ⚠️ Error handling for non-existent notes
 
-**12. read_content Tests (Enhanced):**
+**13. read_content Tests (Enhanced):**
 - ✅ Raw file content access
 - ✅ Binary file handling
 - ✅ Image file reading
 - ⚠️ Large file performance
 
-**13. delete_note Tests (Enhanced):**
+**14. delete_note Tests (Enhanced):**
 - ✅ Single note deletion
 - ✅ Database consistency after deletion
 - ⚠️ Non-existent note handling
 - ✅ Confirmation of successful deletion
 
-**14. list_directory Tests (Enhanced):**
+**15. list_directory Tests (Enhanced):**
 - ✅ Directory content listing
 - ✅ Depth control and filtering
 - ✅ File name globbing
 - ⚠️ Empty directory handling
 
-**15. delete_project Tests (Enhanced):**
+**16. set_default_project Tests (Enhanced):**
+- ✅ Change default project
+- ✅ Configuration persistence
+- ⚠️ Invalid project handling
+
+**17. delete_project Tests (Enhanced):**
 - ✅ Project removal from config
 - ✅ Database cleanup
 - ⚠️ Default project protection
@@ -296,7 +269,7 @@ Test essential MCP tools that form the foundation of Basic Memory:
 1. Technical documentation project
 2. Personal recipe collection project
 3. Learning/course notes project
-4. Specify different projects for different operations
+4. Switch contexts during conversation
 5. Cross-reference related concepts
 
 **Content Evolution:**
@@ -308,13 +281,13 @@ Test essential MCP tools that form the foundation of Basic Memory:
 
 ### Phase 6: Specialized Tools Testing (Tier 4)
 
-**16. canvas Tests (Specialized):**
+**18. canvas Tests (Specialized):**
 - ✅ JSON Canvas generation
 - ✅ Node and edge creation
 - ✅ Obsidian compatibility
 - ⚠️ Complex graph handling
 
-**17. MCP Prompts Tests (Specialized):**
+**19. MCP Prompts Tests (Specialized):**
 - ✅ ai_assistant_guide output
 - ✅ continue_conversation functionality
 - ✅ Formatted search results
@@ -409,7 +382,7 @@ permalink: test-session-[phase]-[timestamp]
 ### 📊 Performance Metrics
 - Average write_note time: 0.3s
 - Search with 100+ notes: 0.6s
-- Project parameter overhead: <0.1s
+- Project switch overhead: 0.1s
 - Memory usage: [observed levels]
 
 ## Relations
@@ -429,7 +402,7 @@ permalink: test-session-[phase]-[timestamp]
 - Learning curve and intuitiveness
 
 **System Behavior:**
-- Stateless operation independence
+- Context preservation across operations
 - memory:// URL navigation reliability
 - Multi-step workflow cohesion
 - Edge case graceful handling

.claude/settings.json

@@ -1,3 +0,0 @@
-{
-  "enabledPlugins": {}
-}

.dxtignore

@@ -0,0 +1,81 @@
+# Development files
+.venv/
+.pytest_cache/
+.coverage*
+htmlcov/
+.ruff_cache/
+**/__pycache__/
+*.pyc
+*.pyo
+*.pyd
+.env
+.env.*
+
+# Testing
+test*/
+tests/
+*.test.js
+*.test.py
+pytest.ini
+
+# Build artifacts
+dist/
+build/
+*.egg-info/
+.tox/
+
+# IDE files
+.idea/
+.vscode/
+*.swp
+*.swo
+*~
+
+# Claude
+.claude
+
+# Git files
+.git/
+.gitignore
+
+# Documentation build
+docs/_build/
+
+# Temporary files
+*.tmp
+*.log
+.DS_Store
+Thumbs.db
+
+# DXT package output
+*.dxt
+dxt-bundle/
+
+
+# Lock files
+uv.lock
+package-lock.json
+yarn.lock
+
+# Docker
+Dockerfile
+docker-compose.yml
+.dockerignore
+
+# CI/CD
+.github/
+justfile
+smithery.yaml
+
+# Project files
+.python.version
+CHANGELOG.md
+CITATION.cff
+CLA.md
+CLAUDE.md
+CODE_OF_CONDUCT.md
+CONTRIBUTING.md
+README.md
+SECURITY.md
+docs/
+llms-install.md

.env.example

@@ -1,28 +0,0 @@
-# Basic Memory Environment Variables Example
-# Copy this file to .env and customize as needed
-# Note: .env files are gitignored and should never be committed
-
-# ============================================================================
-# PostgreSQL Test Database Configuration
-# ============================================================================
-# These variables allow you to override the default test database credentials
-# Default values match docker-compose-postgres.yml for local development
-#
-# Only needed if you want to use different credentials or a remote test database
-# By default, tests use: postgresql://basic_memory_user:dev_password@localhost:5433/basic_memory_test
-
-# Full PostgreSQL test database URL (used by tests and migrations)
-# POSTGRES_TEST_URL=postgresql+asyncpg://basic_memory_user:dev_password@localhost:5433/basic_memory_test
-
-# Individual components (used by justfile postgres-reset command)
-# POSTGRES_USER=basic_memory_user
-# POSTGRES_TEST_DB=basic_memory_test
-
-# ============================================================================
-# Production Database Configuration
-# ============================================================================
-# For production use, set these in your deployment environment
-# DO NOT use the test credentials above in production!
-
-# BASIC_MEMORY_DATABASE_BACKEND=postgres  # or "sqlite"
-# BASIC_MEMORY_DATABASE_URL=postgresql+asyncpg://user:password@host:port/database

.github/workflows/claude-code-review.yml

@@ -1,83 +0,0 @@
-name: Claude Code Review
-
-on:
-  pull_request:
-    types: [opened, synchronize]
-    # Optional: Only run on specific file changes
-    # paths:
-    #   - "src/**/*.ts"
-    #   - "src/**/*.tsx"
-    #   - "src/**/*.js"
-    #   - "src/**/*.jsx"
-
-jobs:
-  claude-review:
-    # Only run for organization members and collaborators
-    if: |
-      github.event.pull_request.author_association == 'OWNER' ||
-      github.event.pull_request.author_association == 'MEMBER' ||
-      github.event.pull_request.author_association == 'COLLABORATOR'
-
-    runs-on: ubuntu-latest
-    permissions:
-      contents: read
-      pull-requests: write
-      issues: read
-      id-token: write
-
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v4
-        with:
-          fetch-depth: 1
-
-      - name: Run Claude Code Review
-        id: claude-review
-        uses: anthropics/claude-code-action@v1
-        with:
-          claude_code_oauth_token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
-          github_token: ${{ secrets.GITHUB_TOKEN }}
-          track_progress: true  # Enable visual progress tracking
-          allowed_bots: '*'
-          prompt: |
-            Review this Basic Memory PR against our team checklist:
-
-            ## Code Quality & Standards
-            - [ ] Follows Basic Memory's coding conventions in CLAUDE.md
-            - [ ] Python 3.12+ type annotations and async patterns
-            - [ ] SQLAlchemy 2.0 best practices
-            - [ ] FastAPI and Typer conventions followed
-            - [ ] 100-character line length limit maintained
-            - [ ] No commented-out code blocks
-
-            ## Testing & Documentation
-            - [ ] Unit tests for new functions/methods
-            - [ ] Integration tests for new MCP tools
-            - [ ] Test coverage for edge cases
-            - [ ] **100% test coverage maintained** (use `# pragma: no cover` only for truly hard-to-test code)
-            - [ ] Documentation updated (README, docstrings)
-            - [ ] CLAUDE.md updated if conventions change
-
-            ## Basic Memory Architecture
-            - [ ] MCP tools follow atomic, composable design
-            - [ ] Database changes include Alembic migrations
-            - [ ] Preserves local-first architecture principles
-            - [ ] Knowledge graph operations maintain consistency
-            - [ ] Markdown file handling preserves integrity
-            - [ ] AI-human collaboration patterns followed
-
-            ## Security & Performance
-            - [ ] No hardcoded secrets or credentials
-            - [ ] Input validation for MCP tools
-            - [ ] Proper error handling and logging
-            - [ ] Performance considerations addressed
-            - [ ] No sensitive data in logs or commits
-            
-            ## Compatability
-            - [ ] File path comparisons must be windows compatible
-            - [ ] Avoid using emojis and unicode characters in console and log output
-
-            Read the CLAUDE.md file for detailed project context. For each checklist item, verify if it's satisfied and comment on any that need attention. Use inline comments for specific code issues and post a summary with checklist results.
-
-          # Allow broader tool access for thorough code review
-          claude_args: '--allowed-tools "Bash(gh pr:*),Bash(gh issue:*),Bash(gh api:*),Bash(git log:*),Bash(git show:*),Read,Grep,Glob"'

.github/workflows/claude-issue-triage.yml

@@ -1,71 +0,0 @@
-name: Claude Issue Triage
-
-on:
-  issues:
-    types: [opened]
-
-jobs:
-  triage:
-    runs-on: ubuntu-latest
-    permissions:
-      issues: write
-      id-token: write
-    steps:
-      - name: Checkout repository
-        uses: actions/checkout@v4
-        with:
-          fetch-depth: 1
-
-      - name: Run Claude Issue Triage
-        uses: anthropics/claude-code-action@v1
-        with:
-          claude_code_oauth_token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
-          track_progress: true  # Show triage progress
-          prompt: |
-            Analyze this new Basic Memory issue and perform triage:
-
-            **Issue Analysis:**
-            1. **Type Classification:**
-               - Bug report (code defect)
-               - Feature request (new functionality)
-               - Enhancement (improvement to existing feature)
-               - Documentation (docs improvement)
-               - Question/Support (user help)
-               - MCP tool issue (specific to MCP functionality)
-
-            2. **Priority Assessment:**
-               - Critical: Security issues, data loss, complete breakage
-               - High: Major functionality broken, affects many users
-               - Medium: Minor bugs, usability issues
-               - Low: Nice-to-have improvements, cosmetic issues
-
-            3. **Component Classification:**
-               - CLI commands
-               - MCP tools
-               - Database/sync
-               - Cloud functionality
-               - Documentation
-               - Testing
-
-            4. **Complexity Estimate:**
-               - Simple: Quick fix, documentation update
-               - Medium: Requires some investigation/testing
-               - Complex: Major feature work, architectural changes
-
-            **Actions to Take:**
-            1. Add appropriate labels using: `gh issue edit ${{ github.event.issue.number }} --add-label "label1,label2"`
-            2. Check for duplicates using: `gh search issues`
-            3. If duplicate found, comment mentioning the original issue
-            4. For feature requests, ask clarifying questions if needed
-            5. For bugs, request reproduction steps if missing
-
-            **Available Labels:**
-            - Type: bug, enhancement, feature, documentation, question, mcp-tool
-            - Priority: critical, high, medium, low
-            - Component: cli, mcp, database, cloud, docs, testing
-            - Complexity: simple, medium, complex
-            - Status: needs-reproduction, needs-clarification, duplicate
-
-            Read the issue carefully and provide helpful triage with appropriate labels.
-
-          claude_args: '--allowed-tools "Bash(gh issue:*),Bash(gh search:*),Read"'

.github/workflows/claude.yml

@@ -9,60 +9,106 @@ on:
     types: [opened, assigned]
   pull_request_review:
     types: [submitted]
-  pull_request_target:
-    types: [opened, synchronize]
 
 jobs:
   claude:
     if: |
-      (
-        (github.event_name == 'issue_comment' && contains(github.event.comment.body, '@claude')) ||
-        (github.event_name == 'pull_request_review_comment' && contains(github.event.comment.body, '@claude')) ||
-        (github.event_name == 'pull_request_review' && contains(github.event.review.body, '@claude')) ||
-        (github.event_name == 'issues' && (contains(github.event.issue.body, '@claude') || contains(github.event.issue.title, '@claude'))) ||
-        (github.event_name == 'pull_request_target' && contains(github.event.pull_request.body, '@claude'))
-      ) && (
-        github.event.comment.author_association == 'OWNER' ||
-        github.event.comment.author_association == 'MEMBER' ||
-        github.event.comment.author_association == 'COLLABORATOR' ||
-        github.event.sender.author_association == 'OWNER' ||
-        github.event.sender.author_association == 'MEMBER' ||
-        github.event.sender.author_association == 'COLLABORATOR' ||
-        github.event.pull_request.author_association == 'OWNER' ||
-        github.event.pull_request.author_association == 'MEMBER' ||
-        github.event.pull_request.author_association == 'COLLABORATOR'
-      )
+      (github.event_name == 'issue_comment' && contains(github.event.comment.body, '@claude')) ||
+      (github.event_name == 'pull_request_review_comment' && contains(github.event.comment.body, '@claude')) ||
+      (github.event_name == 'pull_request_review' && contains(github.event.review.body, '@claude')) ||
+      (github.event_name == 'issues' && (contains(github.event.issue.body, '@claude') || contains(github.event.issue.title, '@claude')))
+
     runs-on: ubuntu-latest
     permissions:
       contents: read
       pull-requests: read
       issues: read
       id-token: write
-      actions: read # Required for Claude to read CI results on PRs
     steps:
+      - name: Check user permissions
+        id: check_membership
+        uses: actions/github-script@v7
+        with:
+          script: |
+            let actor;
+            if (context.eventName === 'issue_comment') {
+              actor = context.payload.comment.user.login;
+            } else if (context.eventName === 'pull_request_review_comment') {
+              actor = context.payload.comment.user.login;
+            } else if (context.eventName === 'pull_request_review') {
+              actor = context.payload.review.user.login;
+            } else if (context.eventName === 'issues') {
+              actor = context.payload.issue.user.login;
+            }
+            
+            console.log(`Checking permissions for user: ${actor}`);
+            
+            // List of explicitly allowed users (organization members)
+            const allowedUsers = [
+              'phernandez',
+              'groksrc',
+              'nellins',
+              'bm-claudeai'
+            ];
+            
+            if (allowedUsers.includes(actor)) {
+              console.log(`User ${actor} is in the allowed list`);
+              core.setOutput('is_member', true);
+              return;
+            }
+            
+            // Fallback: Check if user has repository permissions
+            try {
+              const collaboration = await github.rest.repos.getCollaboratorPermissionLevel({
+                owner: context.repo.owner,
+                repo: context.repo.repo,
+                username: actor
+              });
+              
+              const permission = collaboration.data.permission;
+              console.log(`User ${actor} has permission level: ${permission}`);
+              
+              // Allow if user has push access or higher (write, maintain, admin)
+              const allowed = ['write', 'maintain', 'admin'].includes(permission);
+              
+              core.setOutput('is_member', allowed);
+              
+              if (!allowed) {
+                core.notice(`User ${actor} does not have sufficient repository permissions (has: ${permission})`);
+              }
+            } catch (error) {
+              console.log(`Error checking permissions: ${error.message}`);
+              
+              // Final fallback: Check if user is a public member of the organization
+              try {
+                const membership = await github.rest.orgs.getMembershipForUser({
+                  org: 'basicmachines-co',
+                  username: actor
+                });
+                
+                const allowed = membership.data.state === 'active';
+                core.setOutput('is_member', allowed);
+                
+                if (!allowed) {
+                  core.notice(`User ${actor} is not a public member of basicmachines-co organization`);
+                }
+              } catch (membershipError) {
+                console.log(`Error checking organization membership: ${membershipError.message}`);
+                core.setOutput('is_member', false);
+                core.notice(`User ${actor} does not have access to this repository`);
+              }
+            }
+
       - name: Checkout repository
+        if: steps.check_membership.outputs.is_member == 'true'
         uses: actions/checkout@v4
         with:
-          # For pull_request_target, checkout the PR head to review the actual changes
-          ref: ${{ github.event_name == 'pull_request_target' && github.event.pull_request.head.sha || github.sha }}
           fetch-depth: 1
 
       - name: Run Claude Code
+        if: steps.check_membership.outputs.is_member == 'true'
         id: claude
-        uses: anthropics/claude-code-action@v1
+        uses: anthropics/claude-code-action@beta
         with:
-          claude_code_oauth_token: ${{ secrets.CLAUDE_CODE_OAUTH_TOKEN }}
-          track_progress: true  # Enable visual progress tracking
-          
-          # This is an optional setting that allows Claude to read CI results on PRs
-          additional_permissions: |
-            actions: read
-
-          # Optional: Give a custom prompt to Claude. If this is not specified, Claude will perform the instructions specified in the comment that tagged it.
-          # prompt: 'Update the pull request description to include a summary of changes.'
-
-          # Optional: Add claude_args to customize behavior and configuration
-          # See https://github.com/anthropics/claude-code-action/blob/main/docs/usage.md
-          # or https://docs.claude.com/en/docs/claude-code/sdk#command-line for available options
-          # claude_args: '--model claude-opus-4-1-20250805 --allowed-tools Bash(gh pr:*)'
-
+          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
+          allowed_tools: Bash(uv run pytest),Bash(uv run ruff check . --fix),Bash(uv run ruff format .),Bash(uv run pyright),Bash(just test),Bash(just lint),Bash(just format),Bash(just type-check),Bash(just check),Read,Write,Edit,MultiEdit,Glob,Grep,LS, mcp__web_search

.github/workflows/test.yml

@@ -1,37 +1,43 @@
 name: Tests
 
-concurrency:
-  group: bm-ci-${{ github.workflow }}-${{ github.repository }}-${{ github.head_ref || github.ref }}
-  cancel-in-progress: true
-
 on:
   push:
     branches: [ "main" ]
   pull_request:
     branches: [ "main" ]
+  # pull_request_target runs on the BASE of the PR, not the merge result.
+  # It has write permissions and access to secrets.
+  # It's useful for PRs from forks or automated PRs but requires careful use for security reasons.
+  # See: https://docs.github.com/en/actions/using-workflows/events-that-trigger-workflows#pull_request_target
+  pull_request_target:
+    branches: [ "main" ]
 
 jobs:
-  static-checks:
-    name: Static Checks (Python 3.12)
-    timeout-minutes: 20
+  test:
     runs-on: ubuntu-latest
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: [ "3.12" ]
 
     steps:
       - uses: actions/checkout@v4
         with:
           submodules: true
 
-      - name: Set up Python 3.12
+      - name: Set up Python ${{ matrix.python-version }}
         uses: actions/setup-python@v4
         with:
-          python-version: "3.12"
-          cache: "pip"
+          python-version: ${{ matrix.python-version }}
+          cache: 'pip'
 
       - name: Install uv
         run: |
           pip install uv
 
-      - uses: extractions/setup-just@v3
+      - name: Install just
+        run: |
+          curl --proto '=https' --tlsv1.2 -sSf https://just.systems/install.sh | bash -s -- --to /usr/local/bin
 
       - name: Create virtual env
         run: |
@@ -39,229 +45,13 @@ jobs:
 
       - name: Install dependencies
         run: |
-          uv pip install -e ".[dev]"
+          uv pip install -e .[dev]
 
       - name: Run type checks
         run: |
-          just typecheck
-
-      - name: Run linting
-        run: |
-          just lint
-
-  test-sqlite-unit:
-    name: Test SQLite Unit (${{ matrix.os }}, Python ${{ matrix.python-version }})
-    timeout-minutes: 30
-    needs: [static-checks]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - os: ubuntu-latest
-            python-version: "3.12"
-          - os: ubuntu-latest
-            python-version: "3.13"
-          - os: ubuntu-latest
-            python-version: "3.14"
-          - os: windows-latest
-            python-version: "3.12"
-    runs-on: ${{ matrix.os }}
-
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: true
-
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
-        with:
-          python-version: ${{ matrix.python-version }}
-          cache: 'pip'
-
-      - name: Install uv
-        run: |
-          pip install uv
-
-      - uses: extractions/setup-just@v3
-
-      - name: Create virtual env
-        run: |
-          uv venv
-
-      - name: Install dependencies
-        run: |
-          uv pip install -e ".[dev]"
+          just type-check
 
       - name: Run tests
         run: |
-          just test-unit-sqlite
-
-  test-sqlite-integration:
-    name: Test SQLite Integration (${{ matrix.os }}, Python ${{ matrix.python-version }})
-    timeout-minutes: 45
-    needs: [static-checks]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - os: ubuntu-latest
-            python-version: "3.12"
-          - os: ubuntu-latest
-            python-version: "3.13"
-          - os: ubuntu-latest
-            python-version: "3.14"
-          - os: windows-latest
-            python-version: "3.12"
-    runs-on: ${{ matrix.os }}
-
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: true
-
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
-        with:
-          python-version: ${{ matrix.python-version }}
-          cache: 'pip'
-
-      - name: Install uv
-        run: |
-          pip install uv
-
-      - uses: extractions/setup-just@v3
-
-      - name: Create virtual env
-        run: |
-          uv venv
-
-      - name: Install dependencies
-        run: |
-          uv pip install -e ".[dev]"
-
-      - name: Run tests
-        run: |
-          just test-int-sqlite
-
-  test-postgres-unit:
-    name: Test Postgres Unit (Python ${{ matrix.python-version }})
-    timeout-minutes: 30
-    needs: [static-checks]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - python-version: "3.12"
-          - python-version: "3.13"
-          - python-version: "3.14"
-    runs-on: ubuntu-latest
-
-    # Note: No services section needed - testcontainers handles Postgres in Docker
-
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: true
-
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
-        with:
-          python-version: ${{ matrix.python-version }}
-          cache: 'pip'
-
-      - name: Install uv
-        run: |
-          pip install uv
-
-      - uses: extractions/setup-just@v3
-
-      - name: Create virtual env
-        run: |
-          uv venv
-
-      - name: Install dependencies
-        run: |
-          uv pip install -e ".[dev]"
-
-      - name: Run tests
-        run: |
-          just test-unit-postgres
-
-  test-postgres-integration:
-    name: Test Postgres Integration (Python ${{ matrix.python-version }})
-    timeout-minutes: 45
-    needs: [static-checks]
-    strategy:
-      fail-fast: false
-      matrix:
-        include:
-          - python-version: "3.12"
-          - python-version: "3.13"
-          - python-version: "3.14"
-    runs-on: ubuntu-latest
-
-    # Note: No services section needed - testcontainers handles Postgres in Docker
-
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: true
-
-      - name: Set up Python ${{ matrix.python-version }}
-        uses: actions/setup-python@v4
-        with:
-          python-version: ${{ matrix.python-version }}
-          cache: 'pip'
-
-      - name: Install uv
-        run: |
-          pip install uv
-
-      - uses: extractions/setup-just@v3
-
-      - name: Create virtual env
-        run: |
-          uv venv
-
-      - name: Install dependencies
-        run: |
-          uv pip install -e ".[dev]"
-
-      - name: Run tests
-        run: |
-          just test-int-postgres
-
-  test-semantic:
-    name: Test Semantic (Python 3.12)
-    timeout-minutes: 45
-    needs: [static-checks]
-    runs-on: ubuntu-latest
-
-    steps:
-      - uses: actions/checkout@v4
-        with:
-          submodules: true
-
-      - name: Set up Python 3.12
-        uses: actions/setup-python@v4
-        with:
-          python-version: "3.12"
-          cache: "pip"
-
-      - name: Install uv
-        run: |
-          pip install uv
-
-      - uses: extractions/setup-just@v3
-
-      - name: Create virtual env
-        run: |
-          uv venv
-
-      - name: Install dependencies
-        run: |
-          uv pip install -e ".[dev]"
-
-      - name: Run tests
-        run: |
-          just test-semantic
+          uv pip install pytest pytest-cov
+          just test

.gitignore

@@ -1,7 +1,6 @@
 *.py[cod]
 __pycache__/
 .pytest_cache/
-.testmondata*
 .coverage
 htmlcov/
 
@@ -54,7 +53,7 @@ ENV/
 # claude action
 claude-output
 **/.claude/settings.local.json
-.mcp.json
-.mcpregistry_*
-/.testmondata
-.benchmarks/
+
+# DXT package files
+basic-memory.dxt
+dxt-bundle/

.python-version

@@ -1 +1 @@
-3.14
+3.12

AGENTS.md

@@ -1,446 +0,0 @@
-# AGENTS.md - Basic Memory Project Guide
-
-## Project Overview
-
-Basic Memory is a local-first knowledge management system built on the Model Context Protocol (MCP). It enables
-bidirectional communication between LLMs (like Claude) and markdown files, creating a personal knowledge graph that can
-be traversed using links between documents.
-
-## CODEBASE DEVELOPMENT
-
-### Project information
-
-See the [README.md](README.md) file for a project overview.
-
-### Build and Test Commands
-
-- Install: `just install` or `pip install -e ".[dev]"`
-- Run all tests (SQLite + Postgres): `just test`
-- Run all tests against SQLite: `just test-sqlite`
-- Run all tests against Postgres: `just test-postgres` (uses testcontainers)
-- Run unit tests (SQLite): `just test-unit-sqlite`
-- Run unit tests (Postgres): `just test-unit-postgres`
-- Run integration tests (SQLite): `just test-int-sqlite`
-- Run integration tests (Postgres): `just test-int-postgres`
-- Run impacted tests: `just testmon` (pytest-testmon)
-- Run MCP smoke test: `just test-smoke`
-- Fast local loop: `just fast-check`
-- Local consistency check: `just doctor`
-- Generate HTML coverage: `just coverage`
-- Single test: `pytest tests/path/to/test_file.py::test_function_name`
-- Run benchmarks: `pytest test-int/test_sync_performance_benchmark.py -v -m "benchmark and not slow"`
-- Lint: `just lint` or `ruff check . --fix`
-- Type check: `just typecheck` or `uv run pyright`
-- Type check (supplemental): `just typecheck-ty` or `uv run ty check src/`
-- Format: `just format` or `uv run ruff format .`
-- Run all code checks: `just check` (runs lint, format, typecheck, test)
-- Create db migration: `just migration "Your migration message"`
-- Run development MCP Inspector: `just run-inspector`
-
-**Note:** Project requires Python 3.12+ (uses type parameter syntax and `type` aliases introduced in 3.12)
-
-**Postgres Testing:** Uses [testcontainers](https://testcontainers-python.readthedocs.io/) which automatically spins up a Postgres instance in Docker. No manual database setup required - just have Docker running.
-
-**Doctor Note:** `just doctor` runs with a temporary HOME/config so it won't touch your local Basic Memory settings. It leaves temp dirs in `/tmp` (safe to ignore or remove).
-
-**Testmon Note:** When no files have changed, `just testmon` may collect 0 tests. That's expected and means no impacted tests were detected.
-
-### Code/Test/Verify Loop (fast path)
-
-1) **Code:** make changes.
-2) **Test:** `just fast-check` (lint/format/typecheck + impacted tests + MCP smoke).
-3) **Verify:** `just doctor` (end-to-end file ↔ DB loop in a temp project).
-4) **Full gate (when needed):** `just test` or `just check` for SQLite + Postgres.
-
-If testmon is “cold,” the first run may be long. Subsequent runs get much faster.
-
-### Test Structure
-
-- `tests/` - Unit tests for individual components (mocked, fast)
-- `test-int/` - Integration tests for real-world scenarios (no mocks, realistic)
-- Both directories are covered by unified coverage reporting
-- Benchmark tests in `test-int/` are marked with `@pytest.mark.benchmark`
-- Slow tests are marked with `@pytest.mark.slow`
-- Smoke tests are marked with `@pytest.mark.smoke`
-
-### Code Style Guidelines
-
-- Line length: 100 characters max
-- Python 3.12+ with full type annotations (uses type parameters and type aliases)
-- Format with ruff (consistent styling)
-- Import order: standard lib, third-party, local imports
-- Naming: snake_case for functions/variables, PascalCase for classes
-- Prefer async patterns with SQLAlchemy 2.0
-- Use Pydantic v2 for data validation and schemas
-- CLI uses Typer for command structure
-- API uses FastAPI for endpoints
-- Follow the repository pattern for data access
-- Tools communicate to api routers via the httpx ASGI client (in process)
-
-### Code Change Guidelines
-
-- **Full file read before edits**: Before editing any file, read it in full first to ensure complete context; partial reads lead to corrupted edits
-- **Minimize diffs**: Prefer the smallest change that satisfies the request. Avoid unrelated refactors or style rewrites unless necessary for correctness
-- **No speculative getattr**: Never use `getattr(obj, "attr", default)` when unsure about attribute names. Check the class definition or source code first
-- **Fail fast**: Write code with fail-fast logic by default. Do not swallow exceptions with errors or warnings
-- **No fallback logic**: Do not add fallback logic unless explicitly told to and agreed with the user
-- **No guessing**: Do not say "The issue is..." before you actually know what the issue is. Investigate first.
-
-### Literate Programming Style
-
-Code should tell a story. Comments must explain the "why" and narrative flow, not just the "what".
-
-**Section Headers:**
-For files with multiple phases of logic, add section headers so the control flow reads like chapters:
-```python
-# --- Authentication ---
-# ... auth logic ...
-
-# --- Data Validation ---
-# ... validation logic ...
-
-# --- Business Logic ---
-# ... core logic ...
-```
-
-**Decision Point Comments:**
-For conditionals that materially change behavior (gates, fallbacks, retries, feature flags), add comments with:
-- **Trigger**: what condition causes this branch
-- **Why**: the rationale (cost, correctness, UX, determinism)
-- **Outcome**: what changes downstream
-
-```python
-# Trigger: project has no active sync watcher
-# Why: avoid duplicate file system watchers consuming resources
-# Outcome: starts new watcher, registers in active_watchers dict
-if project_id not in active_watchers:
-    start_watcher(project_id)
-```
-
-**Constraint Comments:**
-If code exists because of a constraint (async requirements, rate limits, schema compatibility), explain the constraint near the code:
-```python
-# SQLite requires WAL mode for concurrent read/write access
-connection.execute("PRAGMA journal_mode=WAL")
-```
-
-**What NOT to Comment:**
-Avoid comments that restate obvious code:
-```python
-# Bad - restates code
-counter += 1  # increment counter
-
-# Good - explains why
-counter += 1  # track retries for backoff calculation
-```
-
-### Codebase Architecture
-
-See [docs/ARCHITECTURE.md](docs/ARCHITECTURE.md) for detailed architecture documentation.
-
-**Directory Structure:**
-- `/alembic` - Alembic db migrations
-- `/api` - FastAPI REST endpoints + `container.py` composition root
-- `/cli` - Typer CLI + `container.py` composition root
-- `/deps` - Feature-scoped FastAPI dependencies (config, db, projects, repositories, services, importers)
-- `/importers` - Import functionality for Claude, ChatGPT, and other sources
-- `/markdown` - Markdown parsing and processing
-- `/mcp` - MCP server + `container.py` composition root + `clients/` typed API clients
-- `/models` - SQLAlchemy ORM models
-- `/repository` - Data access layer
-- `/schemas` - Pydantic models for validation
-- `/services` - Business logic layer
-- `/sync` - File synchronization services + `coordinator.py` for lifecycle management
-
-**Composition Roots:**
-Each entrypoint (API, MCP, CLI) has a composition root that:
-- Reads `ConfigManager` (the only place that reads global config)
-- Resolves runtime mode via `RuntimeMode` enum (TEST > CLOUD > LOCAL)
-- Provides dependencies to downstream code explicitly
-
-**Typed API Clients (MCP):**
-MCP tools use typed clients in `mcp/clients/` to communicate with the API:
-- `KnowledgeClient` - Entity CRUD operations
-- `SearchClient` - Search operations
-- `MemoryClient` - Context building
-- `DirectoryClient` - Directory listing
-- `ResourceClient` - Resource reading
-- `ProjectClient` - Project management
-
-Flow: MCP Tool → Typed Client → HTTP API → Router → Service → Repository
-
-### Development Notes
-
-- MCP tools are defined in src/basic_memory/mcp/tools/
-- MCP prompts are defined in src/basic_memory/mcp/prompts/
-- MCP tools should be atomic, composable operations
-- Use `textwrap.dedent()` for multi-line string formatting in prompts and tools
-- MCP Prompts are used to invoke tools and format content with instructions for an LLM
-- Schema changes require Alembic migrations
-- SQLite is used for indexing and full text search, files are source of truth
-- Testing uses pytest with asyncio support (strict mode)
-- Unit tests (`tests/`) use mocks when necessary; integration tests (`test-int/`) use real implementations
-- By default, tests run against SQLite (fast, no Docker needed)
-- Set `BASIC_MEMORY_TEST_POSTGRES=1` to run against Postgres (uses testcontainers - Docker required)
-- Each test runs in a standalone environment with isolated database and tmp_path directory
-- CI runs SQLite and Postgres tests in parallel for faster feedback
-- Performance benchmarks are in `test-int/test_sync_performance_benchmark.py`
-- Use pytest markers: `@pytest.mark.benchmark` for benchmarks, `@pytest.mark.slow` for slow tests
-- **Coverage must stay at 100%**: Write tests for new code. Only use `# pragma: no cover` when tests would require excessive mocking (e.g., TYPE_CHECKING blocks, error handlers that need failure injection, runtime-mode-dependent code paths)
-
-### Async Client Pattern (Important!)
-
-**MCP tools use `get_project_client()` for per-project routing:**
-
-```python
-from basic_memory.mcp.project_context import get_project_client
-
-@mcp.tool()
-async def my_tool(project: str | None = None, context: Context | None = None):
-    async with get_project_client(project, context) as (client, active_project):
-        # client is routed based on project's mode (local ASGI or cloud HTTP)
-        response = await call_get(client, "/path")
-        return response
-```
-
-**CLI commands and non-project-scoped code use `get_client()` directly:**
-
-```python
-from basic_memory.mcp.async_client import get_client
-
-async def my_cli_command():
-    async with get_client() as client:
-        response = await call_get(client, "/path")
-        return response
-
-# Per-project routing (when project name is known):
-async with get_client(project_name="research") as client:
-    ...
-```
-
-**Do NOT use:**
-- ❌ `from basic_memory.mcp.async_client import client` (deprecated module-level client)
-- ❌ Manual auth header management
-- ❌ `inject_auth_header()` (deleted)
-- ❌ Separate `get_client()` + `get_active_project()` in MCP tools (use `get_project_client()` instead)
-
-**Key principles:**
-- Auth happens at client creation, not per-request
-- Proper resource management via context managers
-- Per-project routing: each project can be LOCAL or CLOUD independently
-- Cloud projects use API key (`cloud_api_key` in config) as Bearer token
-- Routing priority: factory injection > force-local > per-project cloud > global cloud > local ASGI
-- Factory pattern enables dependency injection for cloud consolidation
-
-**For cloud app integration:**
-```python
-from basic_memory.mcp import async_client
-
-# Set custom factory before importing tools
-async_client.set_client_factory(your_custom_factory)
-```
-
-See SPEC-16 for full context manager refactor details.
-
-## BASIC MEMORY PRODUCT USAGE
-
-### Knowledge Structure
-
-- Entity: Any concept, document, or idea represented as a markdown file
-- Observation: A categorized fact about an entity (`- [category] content`)
-- Relation: A directional link between entities (`- relation_type [[Target]]`)
-- Frontmatter: YAML metadata at the top of markdown files
-- Knowledge representation follows precise markdown format:
-    - Observations with [category] prefixes
-    - Relations with WikiLinks [[Entity]]
-    - Frontmatter with metadata
-
-### Basic Memory Commands
-
-**Local Commands:**
-- Check sync status: `basic-memory status`
-- Doctor check (file <-> DB loop): `basic-memory doctor`
-- Import from Claude: `basic-memory import claude conversations`
-- Import from ChatGPT: `basic-memory import chatgpt`
-- Import from Memory JSON: `basic-memory import memory-json`
-- Tool access: `basic-memory tool` (provides CLI access to MCP tools)
-    - Continue: `basic-memory tool continue-conversation --topic="search"`
-
-**Project Management:**
-- List projects: `basic-memory project list`
-- Add project: `basic-memory project add "name" ~/path`
-- Project info: `basic-memory project info`
-- Set cloud mode: `basic-memory project set-cloud "name"`
-- Set local mode: `basic-memory project set-local "name"`
-- One-way sync (local -> cloud): `basic-memory project sync`
-- Bidirectional sync: `basic-memory project bisync`
-- Integrity check: `basic-memory project check`
-
-**Cloud Commands (requires subscription):**
-- Authenticate (global): `basic-memory cloud login`
-- Logout (global): `basic-memory cloud logout`
-- Check cloud status: `basic-memory cloud status`
-- Setup cloud sync: `basic-memory cloud setup`
-- Save API key: `basic-memory cloud set-key bmc_...`
-- Create API key: `basic-memory cloud create-key "name"`
-- Manage snapshots: `basic-memory cloud snapshot [create|list|delete|show|browse]`
-- Restore from snapshot: `basic-memory cloud restore <path> --snapshot <id>`
-
-### MCP Capabilities
-
-- Basic Memory exposes these MCP tools to LLMs:
-
-  **Content Management:**
-    - `write_note(title, content, directory, tags)` - Create/update markdown notes with semantic observations and relations
-    - `read_note(identifier, page, page_size)` - Read notes by title, permalink, or memory:// URL with knowledge graph awareness
-    - `read_content(path)` - Read raw file content (text, images, binaries) without knowledge graph processing
-    - `view_note(identifier, page, page_size)` - View notes as formatted artifacts for better readability
-    - `edit_note(identifier, operation, content)` - Edit notes incrementally (append, prepend, find/replace, replace_section)
-    - `move_note(identifier, destination_path, is_directory)` - Move notes or directories to new locations, updating database and maintaining links
-    - `delete_note(identifier, is_directory)` - Delete notes or directories from the knowledge base
-
-  **Knowledge Graph Navigation:**
-    - `build_context(url, depth, timeframe)` - Navigate the knowledge graph via memory:// URLs for conversation continuity
-    - `recent_activity(type, depth, timeframe)` - Get recently updated information with specified timeframe (e.g., "1d", "1 week")
-    - `list_directory(dir_name, depth, file_name_glob)` - Browse directory contents with filtering and depth control
-
-  **Search & Discovery:**
-    - `search_notes(query, page, page_size, search_type, types, entity_types, after_date)` - Full-text search across all content with advanced filtering options
-
-  **Project Management:**
-    - `list_memory_projects()` - List all available projects with their status
-    - `create_memory_project(project_name, project_path, set_default)` - Create new Basic Memory projects
-    - `delete_project(project_name)` - Delete a project from configuration
-
-  **Visualization:**
-    - `canvas(nodes, edges, title, directory)` - Generate Obsidian canvas files for knowledge graph visualization
-
-  **ChatGPT-Compatible Tools:**
-    - `search(query)` - Search across knowledge base (OpenAI actions compatible)
-    - `fetch(id)` - Fetch full content of a search result document
-
-- MCP Prompts for better AI interaction:
-    - `ai_assistant_guide()` - Guidance on effectively using Basic Memory tools for AI assistants
-    - `continue_conversation(topic, timeframe)` - Continue previous conversations with relevant historical context
-    - `search(query, after_date)` - Search with detailed, formatted results for better context understanding
-    - `recent_activity(timeframe)` - View recently changed items with formatted output
-
-### Cloud Features (v0.15.0+)
-
-Basic Memory now supports cloud synchronization and storage (requires active subscription):
-
-**Authentication:**
-- JWT-based authentication with subscription validation
-- Secure session management with token refresh
-- Support for multiple cloud projects
-
-**Bidirectional Sync:**
-- rclone bisync integration for two-way synchronization
-- Conflict resolution and integrity verification
-- Real-time sync with change detection
-- Mount/unmount cloud storage for direct file access
-
-**Cloud Project Management:**
-- Create and manage projects in the cloud
-- Toggle between local and cloud modes
-- Per-project sync configuration
-- Subscription-based access control
-
-**Security & Performance:**
-- Removed .env file loading for improved security
-- .gitignore integration (respects gitignored files)
-- WAL mode for SQLite performance
-- Background relation resolution (non-blocking startup)
-- API performance optimizations (SPEC-11)
-
-**Per-Project Cloud Routing:**
-
-Individual projects can be routed through the cloud while others stay local, using an API key:
-
-```bash
-# Save API key and set project to cloud mode
-basic-memory cloud set-key bmc_abc123...
-basic-memory project set-cloud research    # route through cloud
-basic-memory project set-local research    # revert to local
-```
-
-MCP tools use `get_project_client()` which automatically routes based on the project's mode. Cloud projects use the `cloud_api_key` from config as Bearer token.
-
-**CLI Routing Flags (Global Cloud Mode):**
-
-When global cloud mode is enabled, CLI commands route to the cloud API by default. Use `--local` and `--cloud` flags to override:
-
-```bash
-# Force local routing (ignore cloud mode)
-basic-memory status --local
-basic-memory project list --local
-
-# Force cloud routing (when cloud mode is disabled)
-basic-memory status --cloud
-basic-memory project info my-project --cloud
-```
-
-Key behaviors:
-- The local MCP server (`basic-memory mcp`) automatically uses local routing
-- This allows simultaneous use of local Claude Desktop and cloud-based clients
-- Some commands (like `project default`, `project sync-config`, `project move`) require `--local` in cloud mode since they modify local configuration
-- Environment variable `BASIC_MEMORY_FORCE_LOCAL=true` forces local routing globally
-- Per-project cloud routing via API key works independently of global cloud mode
-
-## AI-Human Collaborative Development
-
-Basic Memory emerged from and enables a new kind of development process that combines human and AI capabilities. Instead
-of using AI just for code generation, we've developed a true collaborative workflow:
-
-1. AI (LLM) writes initial implementation based on specifications and context
-2. Human reviews, runs tests, and commits code with any necessary adjustments
-3. Knowledge persists across conversations using Basic Memory's knowledge graph
-4. Development continues seamlessly across different AI sessions with consistent context
-5. Results improve through iterative collaboration and shared understanding
-
-This approach has allowed us to tackle more complex challenges and build a more robust system than either humans or AI
-could achieve independently.
-
-**Problem-Solving Guidance:**
-- If a solution isn't working after reasonable effort, suggest alternative approaches
-- Don't persist with a problematic library or pattern when better alternatives exist
-- Example: When py-pglite caused cascading test failures, switching to testcontainers-postgres was the right call
-
-## GitHub Integration
-
-Basic Memory has taken AI-Human collaboration to the next level by integrating Claude directly into the development workflow through GitHub:
-
-### GitHub MCP Tools
-
-Using the GitHub Model Context Protocol server, Claude can now:
-
-- **Repository Management**:
-  - View repository files and structure
-  - Read file contents
-  - Create new branches
-  - Create and update files
-
-- **Issue Management**:
-  - Create new issues
-  - Comment on existing issues
-  - Close and update issues
-  - Search across issues
-
-- **Pull Request Workflow**:
-  - Create pull requests
-  - Review code changes
-  - Add comments to PRs
-
-This integration enables Claude to participate as a full team member in the development process, not just as a code generation tool. Claude's GitHub account ([bm-claudeai](https://github.com/bm-claudeai)) is a member of the Basic Machines organization with direct contributor access to the codebase.
-
-### Collaborative Development Process
-
-With GitHub integration, the development workflow includes:
-
-1. **Direct code review** - Claude can analyze PRs and provide detailed feedback
-2. **Contribution tracking** - All of Claude's contributions are properly attributed in the Git history
-3. **Branch management** - Claude can create feature branches for implementations
-4. **Documentation maintenance** - Claude can keep documentation updated as the code evolves
-5. **Code Commits**: ALWAYS sign off commits with `git commit -s`
-
-This level of integration represents a new paradigm in AI-human collaboration, where the AI assistant becomes a full-fledged team member rather than just a tool for generating code snippets.

CLA.md

@@ -1,71 +1,34 @@
-# Contributor License Agreement
+Developer Certificate of Origin
+Version 1.1
+https://developercertificate.org/
 
-## Copyright Assignment and License Grant
+Copyright (C) 2004, 2006 The Linux Foundation and its contributors.
 
-By signing this Contributor License Agreement ("Agreement"), you accept and agree to the following terms and conditions
-for your present and future Contributions submitted
-to Basic Machines LLC. Except for the license granted herein to Basic Machines LLC and recipients of software
-distributed by Basic Machines LLC, you reserve all right,
-title, and interest in and to your Contributions.
+Everyone is permitted to copy and distribute verbatim copies of this
+license document, but changing it is not allowed.
 
-### 1. Definitions
+Developer's Certificate of Origin 1.1
 
-"You" (or "Your") shall mean the copyright owner or legal entity authorized by the copyright owner that is making this
-Agreement with Basic Machines LLC.
+By making a contribution to this project, I certify that:
 
-"Contribution" shall mean any original work of authorship, including any modifications or additions to an existing work,
-that is intentionally submitted by You to Basic
-Machines LLC for inclusion in, or documentation of, any of the products owned or managed by Basic Machines LLC (the "
-Work").
+(a) The contribution was created in whole or in part by me and I
+have the right to submit it under the open source license
+indicated in the file; or
 
-### 2. Grant of Copyright License
+(b) The contribution is based upon previous work that, to the best
+of my knowledge, is covered under an appropriate open source
+license and I have the right under that license to submit that
+work with modifications, whether created in whole or in part
+by me, under the same open source license (unless I am
+permitted to submit under a different license), as indicated
+in the file; or
 
-Subject to the terms and conditions of this Agreement, You hereby grant to Basic Machines LLC and to recipients of
-software distributed by Basic Machines LLC a perpetual,
-worldwide, non-exclusive, no-charge, royalty-free, irrevocable copyright license to use, copy, modify, merge, publish,
-distribute, sublicense, and/or sell copies of the
-Work, and to permit persons to whom the Work is furnished to do so.
+(c) The contribution was provided directly to me by some other
+person who certified (a), (b) or (c) and I have not modified
+it.
 
-### 3. Assignment of Copyright
-
-You hereby assign to Basic Machines LLC all right, title, and interest worldwide in all Copyright covering your
-Contributions. Basic Machines LLC may license the
-Contributions under any license terms, including copyleft, permissive, commercial, or proprietary licenses.
-
-### 4. Grant of Patent License
-
-Subject to the terms and conditions of this Agreement, You hereby grant to Basic Machines LLC and to recipients of
-software distributed by Basic Machines LLC a perpetual,
-worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to
-make, have made, use, offer to sell, sell, import, and
-otherwise transfer the Work.
-
-### 5. Developer Certificate of Origin
-
-By making a Contribution to this project, You certify that:
-
-(a) The Contribution was created in whole or in part by You and You have the right to submit it under this Agreement; or
-
-(b) The Contribution is based upon previous work that, to the best of Your knowledge, is covered under an appropriate
-open source license and You have the right under that
-license to submit that work with modifications, whether created in whole or in part by You, under this Agreement; or
-
-(c) The Contribution was provided directly to You by some other person who certified (a), (b) or (c) and You have not
-modified it.
-
-(d) You understand and agree that this project and the Contribution are public and that a record of the Contribution (
-including all personal information You submit with
-it, including Your sign-off) is maintained indefinitely and may be redistributed consistent with this project or the
-open source license(s) involved.
-
-### 6. Representations
-
-You represent that you are legally entitled to grant the above license and assignment. If your employer(s) has rights to
-intellectual property that you create that
-includes your Contributions, you represent that you have received permission to make Contributions on behalf of that
-employer, or that your employer has waived such rights
-for your Contributions to Basic Machines LLC.
-
----
-
-This Agreement is effective as of the date you first submit a Contribution to Basic Machines LLC.
+(d) I understand and agree that this project and the contribution
+are public and that a record of the contribution (including all
+personal information I submit with it, including my sign-off) is
+maintained indefinitely and may be redistributed consistent with
+this project or the open source license(s) involved.

CLAUDE.md

@@ -1 +0,0 @@
-AGENTS.md

CLAUDE.md

@@ -0,0 +1,257 @@
+# CLAUDE.md - Basic Memory Project Guide
+
+## Project Overview
+
+Basic Memory is a local-first knowledge management system built on the Model Context Protocol (MCP). It enables
+bidirectional communication between LLMs (like Claude) and markdown files, creating a personal knowledge graph that can
+be traversed using links between documents.
+
+## CODEBASE DEVELOPMENT
+
+### Project information
+
+See the [README.md](README.md) file for a project overview.
+
+### Build and Test Commands
+
+- Install: `just install` or `pip install -e ".[dev]"`
+- Run tests: `uv run pytest -p pytest_mock -v` or `just test`
+- Single test: `pytest tests/path/to/test_file.py::test_function_name`
+- Lint: `just lint` or `ruff check . --fix`
+- Type check: `just type-check` or `uv run pyright`
+- Format: `just format` or `uv run ruff format .`
+- Run all code checks: `just check` (runs lint, format, type-check, test)
+- Create db migration: `just migration "Your migration message"`
+- Run development MCP Inspector: `just run-inspector`
+
+### Code Style Guidelines
+
+- Line length: 100 characters max
+- Python 3.12+ with full type annotations
+- Format with ruff (consistent styling)
+- Import order: standard lib, third-party, local imports
+- Naming: snake_case for functions/variables, PascalCase for classes
+- Prefer async patterns with SQLAlchemy 2.0
+- Use Pydantic v2 for data validation and schemas
+- CLI uses Typer for command structure
+- API uses FastAPI for endpoints
+- Follow the repository pattern for data access
+- Tools communicate to api routers via the httpx ASGI client (in process)
+- avoid using "private" functions in modules or classes (prepended with _)
+
+### Codebase Architecture
+
+- `/alembic` - Alembic db migrations
+- `/api` - FastAPI implementation of REST endpoints
+- `/cli` - Typer command-line interface
+- `/markdown` - Markdown parsing and processing
+- `/mcp` - Model Context Protocol server implementation
+- `/models` - SQLAlchemy ORM models
+- `/repository` - Data access layer
+- `/schemas` - Pydantic models for validation
+- `/services` - Business logic layer
+- `/sync` - File synchronization services
+
+### Development Notes
+
+- MCP tools are defined in src/basic_memory/mcp/tools/
+- MCP prompts are defined in src/basic_memory/mcp/prompts/
+- MCP tools should be atomic, composable operations
+- Use `textwrap.dedent()` for multi-line string formatting in prompts and tools
+- MCP Prompts are used to invoke tools and format content with instructions for an LLM
+- Schema changes require Alembic migrations
+- SQLite is used for indexing and full text search, files are source of truth
+- Testing uses pytest with asyncio support (strict mode)
+- Test database uses in-memory SQLite
+- Avoid creating mocks in tests in most circumstances.
+- Each test runs in a standalone environment with in memory SQLite and tmp_file directory
+- Do not use mocks in tests if possible. Tests run with an in memory sqlite db, so they are not needed. See fixtures in conftest.py
+
+## BASIC MEMORY PRODUCT USAGE
+
+### Knowledge Structure
+
+- Entity: Any concept, document, or idea represented as a markdown file
+- Observation: A categorized fact about an entity (`- [category] content`)
+- Relation: A directional link between entities (`- relation_type [[Target]]`)
+- Frontmatter: YAML metadata at the top of markdown files
+- Knowledge representation follows precise markdown format:
+    - Observations with [category] prefixes
+    - Relations with WikiLinks [[Entity]]
+    - Frontmatter with metadata
+
+### Basic Memory Commands
+
+- Sync knowledge: `basic-memory sync` or `basic-memory sync --watch`
+- Import from Claude: `basic-memory import claude conversations`
+- Import from ChatGPT: `basic-memory import chatgpt`
+- Import from Memory JSON: `basic-memory import memory-json`
+- Check sync status: `basic-memory status`
+- Tool access: `basic-memory tools` (provides CLI access to MCP tools)
+    - Guide: `basic-memory tools basic-memory-guide`
+    - Continue: `basic-memory tools continue-conversation --topic="search"`
+
+### MCP Capabilities
+
+- Basic Memory exposes these MCP tools to LLMs:
+
+  **Content Management:**
+    - `write_note(title, content, folder, tags)` - Create/update markdown notes with semantic observations and relations
+    - `read_note(identifier, page, page_size)` - Read notes by title, permalink, or memory:// URL with knowledge graph awareness
+    - `edit_note(identifier, operation, content)` - Edit notes incrementally (append, prepend, find/replace, section replace)
+    - `move_note(identifier, destination_path)` - Move notes with database consistency and search reindexing
+    - `view_note(identifier)` - Display notes as formatted artifacts for better readability in Claude Desktop
+    - `read_content(path)` - Read raw file content (text, images, binaries) without knowledge graph processing
+    - `delete_note(identifier)` - Delete notes from knowledge base
+
+  **Project Management:**
+    - `list_memory_projects()` - List all available projects with status indicators
+    - `switch_project(project_name)` - Switch to different project context during conversations
+    - `get_current_project()` - Show currently active project with statistics
+    - `create_memory_project(name, path, set_default)` - Create new Basic Memory projects
+    - `delete_project(name)` - Delete projects from configuration and database
+    - `set_default_project(name)` - Set default project in config
+    - `sync_status()` - Check file synchronization status and background operations
+
+  **Knowledge Graph Navigation:**
+    - `build_context(url, depth, timeframe)` - Navigate the knowledge graph via memory:// URLs for conversation continuity
+    - `recent_activity(type, depth, timeframe)` - Get recently updated information with specified timeframe (e.g., "1d", "1 week")
+    - `list_directory(dir_name, depth, file_name_glob)` - List directory contents with filtering and depth control
+
+  **Search & Discovery:**
+    - `search_notes(query, page, page_size)` - Full-text search across all content with filtering options
+
+  **Visualization:**
+    - `canvas(nodes, edges, title, folder)` - Generate Obsidian canvas files for knowledge graph visualization
+
+- MCP Prompts for better AI interaction:
+    - `ai_assistant_guide()` - Guidance on effectively using Basic Memory tools for AI assistants
+    - `continue_conversation(topic, timeframe)` - Continue previous conversations with relevant historical context
+    - `search_notes(query, after_date)` - Search with detailed, formatted results for better context understanding
+    - `recent_activity(timeframe)` - View recently changed items with formatted output
+    - `json_canvas_spec()` - Full JSON Canvas specification for Obsidian visualization
+
+## AI-Human Collaborative Development
+
+Basic Memory emerged from and enables a new kind of development process that combines human and AI capabilities. Instead
+of using AI just for code generation, we've developed a true collaborative workflow:
+
+1. AI (LLM) writes initial implementation based on specifications and context
+2. Human reviews, runs tests, and commits code with any necessary adjustments
+3. Knowledge persists across conversations using Basic Memory's knowledge graph
+4. Development continues seamlessly across different AI sessions with consistent context
+5. Results improve through iterative collaboration and shared understanding
+
+This approach has allowed us to tackle more complex challenges and build a more robust system than either humans or AI
+could achieve independently.
+
+## GitHub Integration
+
+Basic Memory uses Claude directly into the development workflow through GitHub:
+
+### GitHub MCP Tools
+
+Using the GitHub Model Context Protocol server, Claude can:
+
+- **Repository Management**:
+    - View repository files and structure
+    - Read file contents
+    - Create new branches
+    - Create and update files
+
+- **Issue Management**:
+    - Create new issues
+    - Comment on existing issues
+    - Close and update issues
+    - Search across issues
+
+- **Pull Request Workflow**:
+    - Create pull requests
+    - Review code changes
+    - Add comments to PRs
+
+This integration enables Claude to participate as a full team member in the development process, not just as a code
+generation tool. Claude's GitHub account ([bm-claudeai](https://github.com/bm-claudeai)) is a member of the Basic
+Machines organization with direct contributor access to the codebase.
+
+### Collaborative Development Process
+
+With GitHub integration, the development workflow includes:
+
+1. **Direct code review** - Claude can analyze PRs and provide detailed feedback
+2. **Contribution tracking** - All of Claude's contributions are properly attributed in the Git history
+3. **Branch management** - Claude can create feature branches for implementations
+4. **Documentation maintenance** - Claude can keep documentation updated as the code evolves
+
+With this integration, the AI assistant is a full-fledged team member rather than just a tool for generating code
+snippets.
+
+
+### Basic Memory Pro
+
+Basic Memory Pro is a desktop GUI application that wraps the basic-memory CLI/MCP tools:
+
+- Built with Tauri (Rust), React (TypeScript), and a Python FastAPI sidecar
+- Provides visual knowledge graph exploration and project management
+- Uses the same core codebase but adds a desktop-friendly interface
+- Project configuration is shared between CLI and Pro versions
+- Multiple project support with visual switching interface
+
+local repo: /Users/phernandez/dev/basicmachines/basic-memory-pro
+github: https://github.com/basicmachines-co/basic-memory-pro
+
+## Release and Version Management
+
+Basic Memory uses `uv-dynamic-versioning` for automatic version management based on git tags:
+
+### Version Types
+- **Development versions**: Automatically generated from commits (e.g., `0.12.4.dev26+468a22f`)
+- **Beta releases**: Created by tagging with beta suffixes (e.g., `v0.13.0b1`, `v0.13.0rc1`)
+- **Stable releases**: Created by tagging with version numbers (e.g., `v0.13.0`)
+
+### Release Workflows
+
+#### Development Builds (Automatic)
+- Triggered on every push to `main` branch
+- Publishes dev versions like `0.12.4.dev26+468a22f` to PyPI
+- Allows continuous testing of latest changes
+- Users install with: `pip install basic-memory --pre --force-reinstall`
+
+#### Beta/RC Releases (Manual)
+- Create beta tag: `git tag v0.13.0b1 && git push origin v0.13.0b1`
+- Automatically builds and publishes to PyPI as pre-release
+- Users install with: `pip install basic-memory --pre`
+- Use for milestone testing before stable release
+
+#### Stable Releases (Automated)
+- Use the automated release system: `just release v0.13.0`
+- Includes comprehensive quality checks (lint, format, type-check, tests)
+- Automatically updates version in `__init__.py`
+- Creates git tag and pushes to GitHub
+- Triggers GitHub Actions workflow for:
+  - PyPI publication
+  - Homebrew formula update (requires HOMEBREW_TOKEN secret)
+
+**Manual method (legacy):**
+- Create version tag: `git tag v0.13.0 && git push origin v0.13.0`
+
+#### Homebrew Formula Updates
+- Automatically triggered after successful PyPI release for **stable releases only**
+- **Stable releases** (e.g., v0.13.7) automatically update the main `basic-memory` formula
+- **Pre-releases** (dev/beta/rc) are NOT automatically updated - users must specify version manually
+- Updates formula in `basicmachines-co/homebrew-basic-memory` repo
+- Requires `HOMEBREW_TOKEN` secret in GitHub repository settings:
+  - Create a fine-grained Personal Access Token with `Contents: Read and Write` and `Actions: Read` scopes on `basicmachines-co/homebrew-basic-memory`
+  - Add as repository secret named `HOMEBREW_TOKEN` in `basicmachines-co/basic-memory`
+- Formula updates include new version URL and SHA256 checksum
+
+### For Development
+- **Automated releases**: Use `just release v0.13.x` for stable releases and `just beta v0.13.0b1` for beta releases
+- **Quality gates**: All releases require passing lint, format, type-check, and test suites
+- **Version management**: Versions automatically derived from git tags via `uv-dynamic-versioning`
+- **Configuration**: `pyproject.toml` uses `dynamic = ["version"]`
+- **Release automation**: `__init__.py` updated automatically during release process
+- **CI/CD**: GitHub Actions handles building and PyPI publication
+
+## Development Notes
+- make sure you sign off on commits

CONTRIBUTING.md

@@ -27,25 +27,13 @@ project and how to get started as a developer.
 
    > **Note**: Basic Memory uses [just](https://just.systems) as a modern command runner. Install with `brew install just` or `cargo install just`.
 
-3. **Activate the Virtual Environment**
+3. **Run the Tests**:
    ```bash
-   source .venv/bin/activate
-   ```
-
-4. **Run the Tests**:
-   ```bash
-   # Run all tests with unified coverage (unit + integration)
+   # Run all tests
    just test
-
-   # Run unit tests only (fast, no coverage)
-   just test-unit
-
-   # Run integration tests only (fast, no coverage)
-   just test-int
-
-   # Generate HTML coverage report
-   just coverage
-
+   # or
+   uv run pytest -p pytest_mock -v
+   
    # Run a specific test
    pytest tests/path/to/test_file.py::test_function_name
    ```
@@ -141,7 +129,7 @@ agreement to the DCO.
 
 ## Code Style Guidelines
 
-- **Python Version**: Python 3.12+ with full type annotations (3.12+ required for type parameter syntax)
+- **Python Version**: Python 3.12+ with full type annotations
 - **Line Length**: 100 characters maximum
 - **Formatting**: Use ruff for consistent styling
 - **Import Order**: Standard lib, third-party, local imports
@@ -151,78 +139,12 @@ agreement to the DCO.
 
 ## Testing Guidelines
 
-### Test Structure
-
-Basic Memory uses two test directories with unified coverage reporting:
-
-- **`tests/`**: Unit tests that test individual components in isolation
-  - Fast execution with extensive mocking
-  - Test individual functions, classes, and modules
-  - Run with: `just test-unit` (no coverage, fast)
-
-- **`test-int/`**: Integration tests that test real-world scenarios
-  - Test full workflows with real database and file operations
-  - Include performance benchmarks
-  - More realistic but slower than unit tests
-  - Run with: `just test-int` (no coverage, fast)
-
-### Running Tests
-
-```bash
-# Run all tests with unified coverage report
-just test
-
-# Run only unit tests (fast iteration)
-just test-unit
-
-# Run only integration tests
-just test-int
-
-# Generate HTML coverage report
-just coverage
-
-# Run specific test
-pytest tests/path/to/test_file.py::test_function_name
-
-# Run tests excluding benchmarks
-pytest -m "not benchmark"
-
-# Run only benchmark tests
-pytest -m benchmark test-int/test_sync_performance_benchmark.py
-```
-
-### Performance Benchmarks
-
-The `test-int/test_sync_performance_benchmark.py` file contains performance benchmarks that measure sync and indexing speed:
-
-- `test_benchmark_sync_100_files` - Small repository performance
-- `test_benchmark_sync_500_files` - Medium repository performance
-- `test_benchmark_sync_1000_files` - Large repository performance (marked slow)
-- `test_benchmark_resync_no_changes` - Re-sync performance baseline
-
-Run benchmarks with:
-```bash
-# Run all benchmarks (excluding slow ones)
-pytest test-int/test_sync_performance_benchmark.py -v -m "benchmark and not slow"
-
-# Run all benchmarks including slow ones
-pytest test-int/test_sync_performance_benchmark.py -v -m benchmark
-
-# Run specific benchmark
-pytest test-int/test_sync_performance_benchmark.py::test_benchmark_sync_100_files -v
-```
-
-See `test-int/BENCHMARKS.md` for detailed benchmark documentation.
-
-### Testing Best Practices
-
-- **Coverage Target**: We aim for high test coverage for all code
+- **Coverage Target**: We aim for 100% test coverage for all code
 - **Test Framework**: Use pytest for unit and integration tests
-- **Mocking**: Avoid mocking in integration tests; use sparingly in unit tests
+- **Mocking**: Use pytest-mock for mocking dependencies only when necessary
 - **Edge Cases**: Test both normal operation and edge cases
 - **Database Testing**: Use in-memory SQLite for testing database operations
 - **Fixtures**: Use async pytest fixtures for setup and teardown
-- **Markers**: Use `@pytest.mark.benchmark` for benchmarks, `@pytest.mark.slow` for slow tests
 
 ## Release Process
 

Dockerfile

@@ -1,46 +1,26 @@
 FROM python:3.12-slim-bookworm
 
-# Build arguments for user ID and group ID (defaults to 1000)
-ARG UID=1000
-ARG GID=1000
-
 # Copy uv from official image
 COPY --from=ghcr.io/astral-sh/uv:latest /uv /uvx /bin/
 
 # Set environment variables
-# UV_PYTHON_INSTALL_DIR ensures Python is installed to a persistent location
-# that survives in the final image (not in /root/.local which gets lost)
-# UV_PYTHON_PREFERENCE=only-managed tells uv to use its managed Python version
 ENV PYTHONUNBUFFERED=1 \
-    PYTHONDONTWRITEBYTECODE=1 \
-    UV_PYTHON_INSTALL_DIR=/python \
-    UV_PYTHON_PREFERENCE=only-managed
-
-# Create a group and user with the provided UID/GID
-# Check if the GID already exists, if not create appgroup
-RUN (getent group ${GID} || groupadd --gid ${GID} appgroup) && \
-    useradd --uid ${UID} --gid ${GID} --create-home --shell /bin/bash appuser
+    PYTHONDONTWRITEBYTECODE=1
 
 # Copy the project into the image
 ADD . /app
 
-# Install Python 3.13 explicitly and sync the project
+# Sync the project into a new environment, asserting the lockfile is up to date
 WORKDIR /app
-RUN uv python install 3.13
-RUN uv sync --locked --python 3.13
+RUN uv sync --locked
 
-# Create necessary directories and set ownership
-RUN mkdir -p /app/data/basic-memory /app/.basic-memory && \
-    chown -R appuser:${GID} /app
+# Create data directory
+RUN mkdir -p /app/data
 
 # Set default data directory and add venv to PATH
-ENV BASIC_MEMORY_HOME=/app/data/basic-memory \
-    BASIC_MEMORY_PROJECT_ROOT=/app/data \
+ENV BASIC_MEMORY_HOME=/app/data \
     PATH="/app/.venv/bin:$PATH"
 
-# Switch to the non-root user
-USER appuser
-
 # Expose port
 EXPOSE 8000
 
@@ -49,4 +29,4 @@ HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
     CMD basic-memory --version || exit 1
 
 # Use the basic-memory entrypoint to run the MCP server with default SSE transport
-CMD ["basic-memory", "mcp", "--transport", "sse", "--host", "0.0.0.0", "--port", "8000"]
+CMD ["basic-memory", "mcp", "--transport", "sse", "--host", "0.0.0.0", "--port", "8000"]

NOTE-FORMAT.md

@@ -1,494 +0,0 @@
-# Note Format Reference
-
-Every document in Basic Memory is a plain Markdown file. Files are the source of truth — changes to files automatically update the knowledge graph in the database. You maintain complete ownership, files work with git, and knowledge persists independently of any AI conversation.
-
-## Document Structure
-
-A note has three parts: YAML frontmatter, content (observations), and relations.
-
-```markdown
----
-title: Coffee Brewing Methods
-type: note
-tags: [coffee, brewing]
-permalink: coffee-brewing-methods
----
-
-# Coffee Brewing Methods
-
-## Observations
-- [method] Pour over provides more flavor clarity than French press
-- [technique] Water temperature at 205°F extracts optimal compounds #brewing
-- [preference] Ethiopian beans work well with lighter roasts (personal experience)
-
-## Relations
-- relates_to [[Coffee Bean Origins]]
-- requires [[Proper Grinding Technique]]
-- contrasts_with [[Tea Brewing Methods]]
-```
-
-The `## Observations` and `## Relations` headings are conventional but not required — the parser detects observations and relations by their syntax patterns anywhere in the document.
-
-## Frontmatter
-
-YAML metadata between `---` fences at the top of the file.
-
-| Field | Required | Default | Description |
-|-------|----------|---------|-------------|
-| `title` | No | filename stem | Used for linking and references. Auto-set from filename if missing. |
-| `type` | No | `note` | Entity type. Used for schema resolution and filtering. |
-| `tags` | No | `[]` | List or comma-separated string. Used for organization and search. |
-| `permalink` | No | generated from title | Stable identifier. Persists even if the file moves. |
-| `schema` | No | none | Schema attachment — dict (inline), string (reference), or omitted (implicit). |
-
-Custom fields are allowed. Any key not in the standard set is stored as `entity_metadata` and indexed for search and filtering.
-
-```yaml
----
-title: Paul Graham
-type: Person
-tags: [startups, essays, lisp]
-permalink: paul-graham
-status: active
-source: wikipedia
----
-```
-
-Here `status` and `source` are custom fields stored in `entity_metadata`.
-
-### Frontmatter Value Handling
-
-YAML automatically converts some values to native types. Basic Memory normalizes them:
-
-- Date strings (`2025-10-24`) → kept as ISO format strings
-- Numbers (`1.0`) → converted to strings
-- Booleans (`true`) → converted to strings (`"True"`)
-- Lists and dicts → preserved, items normalized recursively
-
-This prevents errors when downstream code expects string values.
-
-## Observations
-
-An observation is a categorized fact about the entity. Written as a Markdown list item.
-
-**Syntax:**
-
-```
-- [category] content text #tag1 #tag2 (context)
-```
-
-| Part | Required | Description |
-|------|----------|-------------|
-| `[category]` | Yes | Classification in square brackets. Any text except `[]()` chars. |
-| content | Yes | The fact or statement. |
-| `#tags` | No | Inline tags. Space-separated, each starting with `#`. |
-| `(context)` | No | Parenthesized text at end of line. Supporting details or source. |
-
-### Examples
-
-```markdown
-- [tech] Uses SQLite for storage #database
-- [design] Follows local-first architecture #architecture
-- [decision] Selected bcrypt for passwords #security (based on OWASP audit)
-- [name] Paul Graham
-- [expertise] Startups
-- [expertise] Lisp
-- [expertise] Essay writing
-```
-
-Array-like fields use repeated categories — multiple `[expertise]` observations above.
-
-### What Is Not an Observation
-
-The parser excludes these list item patterns:
-
-| Pattern | Example | Reason |
-|---------|---------|--------|
-| Checkboxes | `- [ ] Todo item`, `- [x] Done`, `- [-] Cancelled` | Task list syntax |
-| Markdown links | `- [text](url)` | URL link syntax |
-| Bare wiki links | `- [[Target]]` | Treated as a relation instead |
-
-A list item with `#tags` but no `[category]` is still parsed — the tags are extracted and the category defaults to `Note`.
-
-## Relations
-
-Relations connect documents to form the knowledge graph. There are two kinds.
-
-### Explicit Relations
-
-Written as list items with a relation type and a `[[wiki link]]` target.
-
-**Syntax:**
-
-```
-- relation_type [[Target Entity]] (context)
-```
-
-| Part | Required | Description |
-|------|----------|-------------|
-| `relation_type` | No | Text before `[[`. Defaults to `relates_to` if omitted. |
-| `[[Target]]` | Yes | Wiki link to the target entity. Matched by title or permalink. |
-| `(context)` | No | Parenthesized text after `]]`. Supporting details. |
-
-### Examples
-
-```markdown
-- implements [[Search Design]]
-- depends_on [[Database Schema]]
-- works_at [[Y Combinator]] (co-founder)
-- [[Some Entity]]
-```
-
-The last example — a bare `[[wiki link]]` in a list item — gets relation type `relates_to`.
-
-Common relation types:
-- `implements`, `depends_on`, `relates_to`, `inspired_by`
-- `extends`, `part_of`, `contains`, `pairs_with`
-- `works_at`, `authored`, `collaborated_with`
-
-Any text works as a relation type. These are conventions, not a fixed set.
-
-### Inline References
-
-Wiki links appearing in regular prose (not as list items) create implicit `links_to` relations.
-
-```markdown
-This builds on [[Core Design]] and uses [[Utility Functions]].
-```
-
-This creates two relations: `links_to [[Core Design]]` and `links_to [[Utility Functions]]`.
-
-### Forward References
-
-Relations can link to entities that don't exist yet. Basic Memory resolves them when the target is created.
-
-## Permalinks and memory:// URLs
-
-Every document has a unique **permalink** — a stable identifier derived from its title. You can set one explicitly in frontmatter, or let the system generate it.
-
-```yaml
-permalink: auth-approaches-2024
-```
-
-Permalinks form the basis of `memory://` URLs:
-
-```
-memory://auth-approaches-2024        # By permalink
-memory://Authentication Approaches   # By title (auto-resolves)
-memory://project/auth-approaches     # By path
-```
-
-Pattern matching is supported:
-
-```
-memory://auth*                       # Starts with "auth"
-memory://*/approaches                # Ends with "approaches"
-memory://project/*/requirements      # Nested wildcard
-```
-
-## Schemas
-
-Schemas declare the expected structure of a note — which observation categories and relation types a well-formed note should have. They use Picoschema, a compact notation from Google's Dotprompt that fits naturally in YAML frontmatter.
-
-### Picoschema Syntax
-
-```yaml
-schema:
-  name: string, full name              # required field with description
-  email?: string, contact email        # ? = optional
-  role?: string, job title
-  works_at?: Organization, employer    # capitalized type = entity reference
-  tags?(array): string, categories     # array of type
-  status?(enum): [active, inactive]    # enum with allowed values
-  metadata?(object):                   # nested object
-    updated_at?: string
-    source?: string
-```
-
-| Notation | Meaning | Example |
-|----------|---------|---------|
-| `field: type` | Required field | `name: string` |
-| `field?: type` | Optional field | `role?: string` |
-| `field(array): type` | Array of values | `expertise(array): string` |
-| `field?(enum): [vals]` | Enum with allowed values | `status?(enum): [active, inactive]` |
-| `field?(object):` | Nested object with sub-fields | `metadata?(object):` |
-| `, description` | Description after comma | `name: string, full name` |
-| `EntityName` | Capitalized type = entity reference | `works_at?: Organization` |
-
-**Scalar types:** `string`, `integer`, `number`, `boolean`, `any`
-
-Any type not in that set whose first letter is uppercase is treated as an entity reference (a relation target).
-
-### Schema-to-Note Mapping
-
-Schemas validate against existing observation/relation syntax. Note authors don't learn new syntax.
-
-| Schema Declaration | Maps To | Example in Note |
-|--------------------|---------|-----------------|
-| `field: string` | Observation `[field] value` | `- [name] Paul Graham` |
-| `field?(array): string` | Multiple `[field]` observations | `- [expertise] Lisp` (repeated) |
-| `field?: EntityType` | Relation `field [[Target]]` | `- works_at [[Y Combinator]]` |
-| `field?(array): EntityType` | Multiple `field` relations | `- authored [[Book]]` (repeated) |
-| `tags` | Frontmatter `tags` array | `tags: [startups, essays]` |
-| `field?(enum): [vals]` | Observation `[field] value` where value is in the set | `- [status] active` |
-
-Observations and relations not covered by the schema are valid — schemas describe a subset, not a straitjacket.
-
-### Schema Attachment
-
-Three ways to attach a schema to a note, resolved in priority order:
-
-**1. Inline schema** — `schema` is a dict in frontmatter:
-
-```yaml
----
-title: Team Standup 2024-01-15
-type: meeting
-schema:
-  attendees(array): string, who was there
-  decisions(array): string, what was decided
-  action_items(array): string, follow-ups
-  blockers?(array): string, anything stuck
----
-```
-
-Good for one-off structured notes or prototyping a schema before extracting it.
-
-**2. Explicit reference** — `schema` is a string naming a schema note:
-
-```yaml
----
-title: Basic Memory
-schema: SoftwareProject
----
-```
-
-or by permalink:
-
-```yaml
----
-title: LLM Memory Patterns
-schema: schema/research-project
----
-```
-
-Use when the note's `type` differs from the schema it should validate against, or when multiple schema variants exist.
-
-**3. Implicit by type** — no `schema` field, resolved by matching `type`:
-
-```yaml
----
-title: Paul Graham
-type: Person
----
-```
-
-The system looks up a schema note where `entity: Person`. If found, it applies. If not, no validation occurs.
-
-**4. No schema** — perfectly fine. Most notes don't need one.
-
-### Schema Notes
-
-A schema is itself a Basic Memory note with `type: schema`. It lives anywhere (though `schema/` is the conventional directory).
-
-```yaml
-# schema/Person.md
----
-title: Person
-type: schema
-entity: Person
-version: 1
-schema:
-  name: string, full name
-  role?: string, job title or position
-  works_at?: Organization, employer
-  expertise?(array): string, areas of knowledge
-  email?: string, contact email
-settings:
-  validation: warn
----
-
-# Person
-
-A human individual in the knowledge graph.
-```
-
-| Field | Required | Description |
-|-------|----------|-------------|
-| `type` | Yes | Must be `schema` |
-| `entity` | Yes | The entity type this schema describes (e.g., `Person`) |
-| `version` | No | Schema version number (default: `1`) |
-| `schema` | Yes | Picoschema dict defining the fields |
-| `settings.validation` | No | Validation mode (default: `warn`) |
-
-Schema notes are regular notes — they show up in search, can have observations and relations, and participate in the knowledge graph.
-
-### Validation Modes
-
-| Mode | Behavior |
-|------|----------|
-| `warn` | Warnings in output, doesn't block (default) |
-| `strict` | Errors that block sync, for CI/CD enforcement |
-| `off` | No validation |
-
-### Validation Output
-
-```
-$ bm schema validate people/ada-lovelace.md
-
-⚠ Person schema validation:
-  - Missing required field: name (expected [name] observation)
-  - Missing optional field: role
-  - Missing optional field: works_at (no relation found)
-
-ℹ Unmatched observations: [fact] ×2, [born] ×1
-ℹ Unmatched relations: collaborated_with
-```
-
-"Unmatched" items are informational — observations and relations the schema doesn't cover.
-
-### Schema Inference
-
-Generate schemas from existing notes by analyzing observation and relation frequency:
-
-```
-$ bm schema infer Person
-
-Analyzing 30 notes with type: Person...
-
-Observations found:
-  [name]        30/30  100%  → name: string
-  [role]        27/30   90%  → role?: string
-  [expertise]   18/30   60%  → expertise?(array): string
-  [email]        8/30   27%  → email?: string
-
-Relations found:
-  works_at      22/30   73%  → works_at?: Organization
-
-Suggested schema:
-  name: string, full name
-  role?: string, job title
-  expertise?(array): string, areas of knowledge
-  email?: string, contact email
-  works_at?: Organization, employer
-
-Save to schema/Person.md? [y/n]
-```
-
-Frequency thresholds:
-- **100% present** → required field
-- **25%+ present** → optional field
-- **Below 25%** → excluded from suggestion
-
-### Schema Drift Detection
-
-Track how usage patterns shift over time:
-
-```
-$ bm schema diff Person
-
-Schema drift detected:
-
-+ expertise: now in 81% of notes (was 12%)
-- department: dropped to 3% of notes
-~ works_at: cardinality changed (one → many)
-
-Update schema? [y/n/review]
-```
-
-## Complete Examples
-
-### Simple Note (No Schema)
-
-```markdown
----
-title: Project Ideas
-type: note
-tags: [ideas, brainstorm]
----
-
-# Project Ideas
-
-## Observations
-- [idea] Build a CLI tool for markdown linting #tooling
-- [idea] Create a recipe knowledge base #cooking
-- [priority] Focus on developer tools first (Q1 goal)
-
-## Relations
-- inspired_by [[Developer Workflow Research]]
-- part_of [[Q1 Planning]]
-```
-
-### Schema-Validated Note
-
-Schema at `schema/Person.md`:
-
-```yaml
----
-title: Person
-type: schema
-entity: Person
-version: 1
-schema:
-  name: string, full name
-  role?: string, job title or position
-  works_at?: Organization, employer
-  expertise?(array): string, areas of knowledge
-  email?: string, contact email
-settings:
-  validation: warn
----
-
-# Person
-
-A human individual in the knowledge graph.
-```
-
-Note at `people/paul-graham.md`:
-
-```markdown
----
-title: Paul Graham
-type: Person
-tags: [startups, essays, lisp]
----
-
-# Paul Graham
-
-## Observations
-- [name] Paul Graham
-- [role] Essayist and investor
-- [expertise] Startups
-- [expertise] Lisp
-- [expertise] Essay writing
-- [fact] Created Viaweb, the first web app
-
-## Relations
-- works_at [[Y Combinator]]
-- authored [[Hackers and Painters]]
-```
-
-The `[fact]` observation and `authored` relation are not in the schema — they're valid, just unmatched. The schema only checks that `[name]` exists (required) and looks for optional fields like `[role]`, `[expertise]`, and `works_at`.
-
-### Inline Schema Note
-
-```markdown
----
-title: Team Standup 2024-01-15
-type: meeting
-schema:
-  attendees(array): string, who was there
-  decisions(array): string, what was decided
-  action_items(array): string, follow-ups
-  blockers?(array): string, anything stuck
----
-
-# Team Standup 2024-01-15
-
-## Observations
-- [attendees] Paul
-- [attendees] Sarah
-- [decisions] Ship v2 by Friday
-- [action_items] Paul to review PR #42
-- [blockers] Waiting on API credentials
-```

README.md

@@ -1,4 +1,3 @@
-<!-- mcp-name: io.github.basicmachines-co/basic-memory -->
 [![License: AGPL v3](https://img.shields.io/badge/License-AGPL_v3-blue.svg)](https://www.gnu.org/licenses/agpl-3.0)
 [![PyPI version](https://badge.fury.io/py/basic-memory.svg)](https://badge.fury.io/py/basic-memory)
 [![Python 3.12+](https://img.shields.io/badge/python-3.12+-blue.svg)](https://www.python.org/downloads/)
@@ -6,16 +5,7 @@
 [![Ruff](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/astral-sh/ruff/main/assets/badge/v2.json)](https://github.com/astral-sh/ruff)
 ![](https://badge.mcpx.dev?type=server 'MCP Server')
 ![](https://badge.mcpx.dev?type=dev 'MCP Dev')
-
-## 🚀 Basic Memory Cloud is Live!
-
-- **Cross-device and multi-platform support is here.** Your knowledge graph now works on desktop, web, and mobile.
-- **Cloud is optional.** The local-first open-source workflow continues as always.
-- **OSS discount:** use code `BMFOSS` for 20% off for 3 months.
-
-[Sign up now →](https://basicmemory.com?utm_source=github&utm_medium=referral&utm_campaign=readme)
-
-with a 7 day free trial
+[![smithery badge](https://smithery.ai/badge/@basicmachines-co/basic-memory)](https://smithery.ai/server/@basicmachines-co/basic-memory)
 
 # Basic Memory
 
@@ -23,21 +13,11 @@ Basic Memory lets you build persistent knowledge through natural conversations w
 Claude, while keeping everything in simple Markdown files on your computer. It uses the Model Context Protocol (MCP) to
 enable any compatible LLM to read and write to your local knowledge base.
 
-## What's New in v0.19.0
-
-- **Semantic Vector Search** — find notes by meaning, not just keywords. Combines full-text and vector similarity for hybrid search with FastEmbed embeddings.
-- **Schema System** — infer, validate, and diff the structure of your knowledge base with `schema_infer`, `schema_validate`, and `schema_diff` tools.
-- **Per-Project Cloud Routing** — route individual projects through the cloud while others stay local, using API key authentication (`basic-memory project set-cloud`).
-- **FastMCP 3.0** — upgraded to FastMCP 3.0 with tool annotations for better client integration.
-- **CLI Overhaul** — JSON output mode (`--json`) for scripting, workspace-aware commands, and an htop-inspired project dashboard.
-- **Smarter Editing** — `edit_note` append/prepend auto-creates notes if they don't exist; `write_note` has an overwrite guard to prevent accidental data loss.
-- **Richer Search Results** — matched chunk text returned in search results for better context.
-
-See the full [CHANGELOG](CHANGELOG.md) for details.
-
-- Website: [basicmemory.com](https://basicmemory.com?utm_source=github&utm_medium=referral&utm_campaign=readme)
-- Documentation: [docs.basicmemory.com](https://docs.basicmemory.com?utm_source=github&utm_medium=referral&utm_campaign=readme)
-- Community: [Discord](https://discord.gg/tyvKNccgqN?utm_source=github&utm_medium=referral&utm_campaign=readme)
+- Website: https://basicmemory.com
+- Company: https://basicmachines.co
+- Documentation: https://memory.basicmachines.co
+- Discord: https://discord.gg/tyvKNccgqN
+- YouTube: https://www.youtube.com/@basicmachines-co
 
 ## Pick up your conversation right where you left off
 
@@ -53,6 +33,10 @@ https://github.com/user-attachments/assets/a55d8238-8dd0-454a-be4c-8860dbbd0ddc
 # Install with uv (recommended)
 uv tool install basic-memory
 
+# or with Homebrew
+brew tap basicmachines-co/basic-memory
+brew install basic-memory
+
 # Configure Claude Desktop (edit ~/Library/Application Support/Claude/claude_desktop_config.json)
 # Add this to your config:
 {
@@ -75,35 +59,22 @@ uv tool install basic-memory
 
 You can view shared context via files in `~/basic-memory` (default directory location).
 
-## Automatic Updates
+### Alternative Installation via Smithery
 
-Basic Memory includes a default-on auto-update flow for CLI installs.
-
-- **Auto-install supported:** `uv tool` and Homebrew installs
-- **Default check interval:** every 24 hours (`86400` seconds)
-- **MCP-safe behavior:** update checks run silently in `basic-memory mcp` mode
-- **`uvx` behavior:** skipped (runtime is ephemeral and managed by `uvx`)
-
-Manual update commands:
+You can use [Smithery](https://smithery.ai/server/@basicmachines-co/basic-memory) to automatically configure Basic
+Memory for Claude Desktop:
 
 ```bash
-# Check now and install if supported
-bm update
-
-# Check only, do not install
-bm update --check
+npx -y @smithery/cli install @basicmachines-co/basic-memory --client claude
 ```
 
-Config options in `~/.basic-memory/config.json`:
+This installs and configures Basic Memory without requiring manual edits to the Claude Desktop configuration file. Note: The Smithery installation uses their hosted MCP server, while your data remains stored locally as Markdown files.
 
-```json
-{
-  "auto_update": true,
-  "update_check_interval": 86400
-}
-```
+### Glama.ai
 
-To disable automatic updates, set `"auto_update": false`.
+<a href="https://glama.ai/mcp/servers/o90kttu9ym">
+  <img width="380" height="200" src="https://glama.ai/mcp/servers/o90kttu9ym/badge" alt="basic-memory MCP server" />
+</a>
 
 ## Why Basic Memory?
 
@@ -135,9 +106,6 @@ With Basic Memory, you can:
 - Keep everything local and under your control
 - Use familiar tools like Obsidian to view and edit notes
 - Build a personal knowledge base that grows over time
-- Sync your knowledge to the cloud with bidirectional synchronization
-- Authenticate and manage cloud projects with subscription validation
-- Mount cloud storage for direct file access
 
 ## How It Works in Practice
 
@@ -191,7 +159,8 @@ The note embeds semantic content and links to other topics via simple Markdown f
 
 3. You see this file on your computer in real time in the current project directory (default `~/$HOME/basic-memory`).
 
-- Realtime sync can be enabled via running `basic-memory sync --watch`
+- Realtime sync is enabled by default starting with v0.12.0
+- Project switching during conversations is supported starting with v0.13.0
 
 4. In a chat with the LLM, you can reference a topic:
 
@@ -301,6 +270,13 @@ Examples of relations:
 ```
 
 ## Using with VS Code
+For one-click installation, click one of the install buttons below...
+
+[![Install with UV in VS Code](https://img.shields.io/badge/VS_Code-UV-0098FF?style=flat-square&logo=visualstudiocode&logoColor=white)](https://insiders.vscode.dev/redirect/mcp/install?name=basic-memory&config=%7B%22command%22%3A%22uvx%22%2C%22args%22%3A%5B%22basic-memory%22%2C%22mcp%22%5D%7D) [![Install with UV in VS Code Insiders](https://img.shields.io/badge/VS_Code_Insiders-UV-24bfa5?style=flat-square&logo=visualstudiocode&logoColor=white)](https://insiders.vscode.dev/redirect/mcp/install?name=basic-memory&config=%7B%22command%22%3A%22uvx%22%2C%22args%22%3A%5B%22basic-memory%22%2C%22mcp%22%5D%7D&quality=insiders)
+
+You can use Basic Memory with VS Code to easily retrieve and store information while coding. Click the installation buttons above for one-click setup, or follow the manual installation instructions below.
+
+### Manual Installation
 
 Add the following JSON block to your User Settings (JSON) file in VS Code. You can do this by pressing `Ctrl + Shift + P` and typing `Preferences: Open User Settings (JSON)`.
 
@@ -330,8 +306,6 @@ Optionally, you can add it to a file called `.vscode/mcp.json` in your workspace
 }
 ```
 
-You can use Basic Memory with VS Code to easily retrieve and store information while coding.
-
 ## Using with Claude Desktop
 
 Basic Memory is built using the MCP (Model Context Protocol) and works with the Claude desktop app (https://claude.ai/):
@@ -355,7 +329,8 @@ for OS X):
 }
 ```
 
-If you want to use a specific project (see [Multiple Projects](#multiple-projects) below), update your Claude Desktop
+If you want to use a specific project (see [Multiple Projects](docs/User%20Guide.md#multiple-projects)), update your
+Claude Desktop
 config:
 
 ```json
@@ -365,9 +340,9 @@ config:
       "command": "uvx",
       "args": [
         "basic-memory",
-        "mcp",
         "--project",
-        "your-project-name"
+        "your-project-name",
+        "mcp"
       ]
     }
   }
@@ -376,134 +351,27 @@ config:
 
 2. Sync your knowledge:
 
-```bash
-# One-time sync of local knowledge updates
-basic-memory sync
+Basic Memory will sync the files in your project in real time if you make manual edits.
+
+3. In Claude Desktop, the LLM can now use these tools:
 
-# Run realtime sync process (recommended)
-basic-memory sync --watch
-```
-
-3. Cloud features (optional, requires subscription):
-
-```bash
-# Authenticate with cloud (stores OAuth token locally)
-basic-memory cloud login
-
-# (Optional) install/configure rclone for file sync commands
-basic-memory cloud setup
-
-# Check cloud auth + health
-basic-memory cloud status
-```
-
-**Per-Project Cloud Routing** (API key based):
-
-Individual projects can be routed through the cloud while others stay local. This uses an API key for routed
-project calls:
-
-```bash
-# Save an API key (create one in the web app or via CLI)
-basic-memory cloud set-key bmc_abc123...
-# Or create one via CLI (requires OAuth login first)
-basic-memory cloud create-key "my-laptop"
-
-# Set a project to route through cloud
-basic-memory project set-cloud research
-
-# Revert a project to local mode
-basic-memory project set-local research
-
-# List projects and route metadata
-basic-memory project list
-```
-
-`basic-memory cloud login` / `basic-memory cloud logout` are authentication commands. They do not change default CLI
-routing behavior.
-
-**Routing Flags**:
-
-Use routing flags to disambiguate command targets:
-
-```bash
-# Force local routing for this command
-basic-memory status --local
-basic-memory project list --local
-basic-memory project ls --name main --local
-
-# Force cloud routing for this command
-basic-memory status --cloud
-basic-memory project info my-project --cloud
-basic-memory project ls --name main --cloud
-```
-
-No-flag behavior defaults to local when no project context is present.
-
-The local MCP server routes per transport: `--transport stdio` honors per-project routing
-(local or cloud), while `--transport streamable-http` and `--transport sse` always route locally.
-
-**CLI Note Editing (`tool edit-note`):**
-
-```bash
-# Append content
-basic-memory tool edit-note project-plan --operation append --content $'\n## Next Steps\n- Finalize rollout'
-
-# Find/replace with replacement count validation
-basic-memory tool edit-note docs/api --operation find_replace --find-text "v0.14.0" --content "v0.15.0" --expected-replacements 2
-
-# Replace a section body
-basic-memory tool edit-note docs/setup --operation replace_section --section "## Installation" --content $'Updated install steps\n- Run just install'
-
-# JSON metadata output for integrations
-basic-memory tool edit-note docs/setup --operation append --content $'\n- Added note' --format json
-```
-
-4. In Claude Desktop, the LLM can now use these tools:
-
-**Content Management:**
-```
-write_note(title, content, folder, tags, output_format="text"|"json") - Create or update notes
-read_note(identifier, page, page_size, output_format="text"|"json") - Read notes by title or permalink
-read_content(path) - Read raw file content (text, images, binaries)
-view_note(identifier) - View notes as formatted artifacts
-edit_note(identifier, operation, content, output_format="text"|"json") - Edit notes incrementally
-move_note(identifier, destination_path, output_format="text"|"json") - Move notes with database consistency
-delete_note(identifier, output_format="text"|"json") - Delete notes from knowledge base
-```
-
-**Knowledge Graph Navigation:**
-```
-build_context(url, depth, timeframe, output_format="json"|"text") - Navigate knowledge graph via memory:// URLs
-recent_activity(type, depth, timeframe, output_format="text"|"json") - Find recently updated information
-list_directory(dir_name, depth) - Browse directory contents with filtering
-```
-
-**Search & Discovery:**
-```
-search(query, page, page_size) - Search across your knowledge base
-search_notes(query, page, page_size, search_type, types, entity_types, after_date, metadata_filters, tags, status, project) - Search with filters (query is optional for filter-only searches)
-```
-
-**Project Management:**
-```
-list_memory_projects(output_format="text"|"json") - List all available projects
-create_memory_project(project_name, project_path, output_format="text"|"json") - Create new projects
-get_current_project() - Show current project stats
-sync_status() - Check synchronization status
-```
-
-`output_format` defaults to `"text"` for these tools, preserving current human-readable responses.
-`build_context` defaults to `"json"` and can be switched to `"text"` when compact markdown output is preferred.
-
-**Cloud Discovery (opt-in):**
-```
-cloud_info() - Show optional Cloud overview and setup guidance
-release_notes() - Show latest release notes
-```
-
-**Visualization:**
 ```
+write_note(title, content, folder, tags) - Create or update notes
+read_note(identifier, page, page_size) - Read notes by title or permalink
+edit_note(identifier, operation, content) - Edit notes incrementally (append, prepend, find/replace)
+move_note(identifier, destination_path) - Move notes with database consistency
+view_note(identifier) - Display notes as formatted artifacts for better readability
+build_context(url, depth, timeframe) - Navigate knowledge graph via memory:// URLs
+search_notes(query, page, page_size) - Search across your knowledge base
+recent_activity(type, depth, timeframe) - Find recently updated information
 canvas(nodes, edges, title, folder) - Generate knowledge visualizations
+list_memory_projects() - List all available projects with status
+switch_project(project_name) - Switch to different project context
+get_current_project() - Show current project and statistics
+create_memory_project(name, path, set_default) - Create new projects
+delete_project(name) - Delete projects from configuration
+set_default_project(name) - Set default project
+sync_status() - Check file synchronization status
 ```
 
 5. Example prompts to try:
@@ -514,151 +382,63 @@ canvas(nodes, edges, title, folder) - Generate knowledge visualizations
 "Create a canvas visualization of my project components"
 "Read my notes on the authentication system"
 "What have I been working on in the past week?"
+"Switch to my work-notes project"
+"List all my available projects"
+"Edit my coffee brewing note to add a new technique"
+"Move my old meeting notes to the archive folder"
 ```
 
 ## Futher info
 
-See the [Documentation](https://docs.basicmemory.com?utm_source=github&utm_medium=referral&utm_campaign=readme) for more info, including:
+See the [Documentation](https://memory.basicmachines.co/) for more info, including:
 
-- [Complete User Guide](https://docs.basicmemory.com/user-guide/?utm_source=github&utm_medium=referral&utm_campaign=readme)
-- [CLI tools](https://docs.basicmemory.com/guides/cli-reference/?utm_source=github&utm_medium=referral&utm_campaign=readme)
-- [Cloud CLI and Sync](https://docs.basicmemory.com/guides/cloud-cli/?utm_source=github&utm_medium=referral&utm_campaign=readme)
-- [Managing multiple Projects](https://docs.basicmemory.com/guides/cli-reference/?utm_source=github&utm_medium=referral&utm_campaign=readme#project)
-- [Importing data from OpenAI/Claude Projects](https://docs.basicmemory.com/guides/cli-reference/?utm_source=github&utm_medium=referral&utm_campaign=readme#import)
+- [Complete User Guide](https://memory.basicmachines.co/docs/user-guide)
+- [CLI tools](https://memory.basicmachines.co/docs/cli-reference)
+- [Managing multiple Projects](https://memory.basicmachines.co/docs/cli-reference#project)
+- [Importing data from OpenAI/Claude Projects](https://memory.basicmachines.co/docs/cli-reference#import)
 
-## Telemetry
-
-Basic Memory collects anonymous, minimal usage events to understand how the CLI-to-cloud conversion funnel performs. This helps us prioritize features and improve the product.
-
-**What we collect:**
-- Cloud promo impressions (when the promo banner is shown)
-- Cloud login attempts and outcomes
-- Promo opt-out events
-
-**What we do NOT collect:**
-- No file contents, note titles, or knowledge base data
-- No personally identifiable information (PII)
-- No IP address tracking or fingerprinting
-- No per-command or per-tool-call tracking
-
-Events are sent to our [Umami Cloud](https://umami.is) instance, an open-source, privacy-focused analytics platform. Events are fire-and-forget on a background thread — analytics never blocks or slows the CLI.
-
-**Opt out** by setting the environment variable:
+## Installation Options
 
+### Stable Release
 ```bash
-export BASIC_MEMORY_NO_PROMOS=1
+pip install basic-memory
 ```
 
-This disables both promo messages and all telemetry events.
-
-## Logging
-
-Basic Memory uses [Loguru](https://github.com/Delgan/loguru) for logging. The logging behavior varies by entry point:
-
-| Entry Point | Default Behavior | Use Case |
-|-------------|------------------|----------|
-| CLI commands | File only | Prevents log output from interfering with command output |
-| MCP server | File only | Stdout would corrupt the JSON-RPC protocol |
-| API server | File (local) or stdout (cloud) | Docker/cloud deployments use stdout |
-
-**Log file location:** `~/.basic-memory/basic-memory.log` (10MB rotation, 10 days retention)
-
-### Environment Variables
-
-| Variable | Default | Description |
-|----------|---------|-------------|
-| `BASIC_MEMORY_LOG_LEVEL` | `INFO` | Log level: DEBUG, INFO, WARNING, ERROR |
-| `BASIC_MEMORY_CLOUD_MODE` | `false` | When `true`, API logs to stdout with structured context |
-| `BASIC_MEMORY_FORCE_LOCAL` | `false` | When `true`, forces local API routing |
-| `BASIC_MEMORY_FORCE_CLOUD` | `false` | When `true`, forces cloud API routing |
-| `BASIC_MEMORY_EXPLICIT_ROUTING` | `false` | When `true`, marks route selection as explicit (`--local`/`--cloud`) |
-| `BASIC_MEMORY_ENV` | `dev` | Set to `test` for test mode (stderr only) |
-| `BASIC_MEMORY_NO_PROMOS` | `false` | When `true`, disables cloud promo messages and telemetry |
-
-### Examples
-
+### Beta/Pre-releases
 ```bash
-# Enable debug logging
-BASIC_MEMORY_LOG_LEVEL=DEBUG basic-memory sync
-
-# View logs
-tail -f ~/.basic-memory/basic-memory.log
-
-# Cloud/Docker mode (stdout logging with structured context)
-BASIC_MEMORY_CLOUD_MODE=true uvicorn basic_memory.api.app:app
+pip install basic-memory --pre
 ```
 
-## Development
-
-### Running Tests
-
-Basic Memory supports dual database backends (SQLite and Postgres). By default, tests run against SQLite. Set `BASIC_MEMORY_TEST_POSTGRES=1` to run against Postgres (uses testcontainers - Docker required).
-
-**Quick Start:**
+### Development Builds
+Development versions are automatically published on every commit to main with versions like `0.12.4.dev26+468a22f`:
 ```bash
-# Run all tests against SQLite (default, fast)
-just test-sqlite
-
-# Run all tests against Postgres (uses testcontainers)
-just test-postgres
-
-# Run both SQLite and Postgres tests
-just test
+pip install basic-memory --pre --force-reinstall
 ```
 
-**Available Test Commands:**
+### Docker
 
-- `just test` - Run all tests against both SQLite and Postgres
-- `just test-sqlite` - Run all tests against SQLite (fast, no Docker needed)
-- `just test-postgres` - Run all tests against Postgres (uses testcontainers)
-- `just test-unit-sqlite` - Run unit tests against SQLite
-- `just test-unit-postgres` - Run unit tests against Postgres
-- `just test-int-sqlite` - Run integration tests against SQLite
-- `just test-int-postgres` - Run integration tests against Postgres
-- `just test-windows` - Run Windows-specific tests (auto-skips on other platforms)
-- `just test-benchmark` - Run performance benchmark tests
-- `just testmon` - Run tests impacted by recent changes (pytest-testmon)
-- `just test-smoke` - Run fast MCP end-to-end smoke test
-- `just fast-check` - Run fix/format/typecheck + impacted tests + smoke test
-- `just doctor` - Run local file <-> DB consistency checks with temp config
+Run Basic Memory in a container with volume mounting for your Obsidian vault:
 
-**Postgres Testing:**
-
-Postgres tests use [testcontainers](https://testcontainers-python.readthedocs.io/) which automatically spins up a Postgres instance in Docker. No manual database setup required - just have Docker running.
-
-**Testmon Note:** When no files have changed, `just testmon` may collect 0 tests. That's expected and means no impacted tests were detected.
-
-**Test Markers:**
-
-Tests use pytest markers for selective execution:
-- `windows` - Windows-specific database optimizations
-- `benchmark` - Performance tests (excluded from default runs)
-- `smoke` - Fast MCP end-to-end smoke tests
-
-**Other Development Commands:**
 ```bash
-just install          # Install with dev dependencies
-just lint             # Run linting checks
-just typecheck        # Run type checking
-just typecheck-ty     # Run ty type checking (incremental supplement to pyright)
-just format           # Format code with ruff
-just fast-check       # Fast local loop (fix/format/typecheck + testmon + smoke)
-just doctor           # Local consistency check (temp config)
-just check            # Run all quality checks
-just migration "msg"  # Create database migration
+# Clone and start with Docker Compose
+git clone https://github.com/basicmachines-co/basic-memory.git
+cd basic-memory
+
+# Edit docker-compose.yml to point to your Obsidian vault
+# Then start the container
+docker-compose up -d
 ```
 
-**Type Checking Strategy:**
-- `just typecheck` (Pyright) remains the primary, blocking type checker.
-- `just typecheck-ty` (Astral `ty`) is available as a supplemental checker while rules are adopted incrementally.
-- We recommend running both locally while reducing `ty` diagnostics over time.
-
-**Local Consistency Check:**
+Or use Docker directly:
 ```bash
-basic-memory doctor   # Verifies file <-> database sync in a temp project
+docker run -d \
+  --name basic-memory-server \
+  -v /path/to/your/obsidian-vault:/data/knowledge:rw \
+  -v basic-memory-config:/root/.basic-memory:rw \
+  ghcr.io/basicmachines-co/basic-memory:latest
 ```
 
-See the [justfile](justfile) for the complete list of development commands.
+See [Docker Setup Guide](docs/Docker.md) for detailed configuration options, multiple project setup, and integration examples.
 
 ## License
 
@@ -677,4 +457,4 @@ and submitting PRs.
  </picture>
 </a>
 
-Built with ♥️ by [Basic Machines](https://basicmachines.co?utm_source=github&utm_medium=referral&utm_campaign=readme)
+Built with ♥️ by Basic Machines

docker-compose-postgres.yml

@@ -1,42 +0,0 @@
-# Docker Compose configuration for Basic Memory with PostgreSQL
-# Use this for local development and testing with Postgres backend
-#
-# Usage:
-#   docker-compose -f docker-compose-postgres.yml up -d
-#   docker-compose -f docker-compose-postgres.yml down
-
-services:
-  postgres:
-    image: postgres:17
-    container_name: basic-memory-postgres
-    environment:
-      # Local development/test credentials - NOT for production
-      # These values are referenced by tests and justfile commands
-      POSTGRES_DB: basic_memory
-      POSTGRES_USER: basic_memory_user
-      POSTGRES_PASSWORD: dev_password  # Simple password for local testing only
-    ports:
-      - "5433:5432"
-    volumes:
-      - postgres_data:/var/lib/postgresql/data
-    healthcheck:
-      test: ["CMD-SHELL", "pg_isready -U basic_memory_user -d basic_memory"]
-      interval: 10s
-      timeout: 5s
-      retries: 5
-    restart: unless-stopped
-
-volumes:
-  # Named volume for Postgres data
-  postgres_data:
-    driver: local
-
-  # Named volume for persistent configuration
-  # Database will be stored in Postgres, not in this volume
-  basic-memory-config:
-    driver: local
-
-# Network configuration (optional)
-# networks:
-#   basic-memory-net:
-#     driver: bridge

docs/AI Assistant Guide.md

@@ -0,0 +1,431 @@
+---
+title: AI Assistant Guide
+type: note
+permalink: docs/ai-assistant-guide
+---
+> Note: This is an optional document that can be copy/pasted into the project knowledge for an LLM to provide a full description of how it can work with Basic Memory. It is provided as a helpful resource. The tools contain extensive usage description prompts with enable the LLM to understand them. 
+
+You can [download](https://github.com/basicmachines-co/basic-memory/blob/main/docs/AI%20Assistant%20Guide.md) the contents of this file from GitHub
+# AI Assistant Guide for Basic Memory
+
+This guide helps you, the AI assistant, use Basic Memory tools effectively when working with users. It covers reading, writing, and navigating knowledge through the Model Context Protocol (MCP).
+
+## Quick Reference
+
+**Essential Tools:**
+- `write_note()` - Create/update notes (primary tool)
+- `read_note()` - Read existing content
+- `search_notes()` - Find information
+- `edit_note()` - Modify existing notes incrementally (v0.13.0)
+- `move_note()` - Organize files with database consistency (v0.13.0)
+
+**Project Management (v0.13.0):**
+- `list_projects()` - Show available projects
+- `switch_project()` - Change active project
+- `get_current_project()` - Current project info
+
+**Key Principles:**
+1. **Build connections** - Rich knowledge graphs > isolated notes
+2. **Ask permission** - "Would you like me to record this?"
+3. **Use exact titles** - For accurate `[[WikiLinks]]`
+4. **Leverage v0.13.0** - Edit incrementally, organize proactively, switch projects contextually
+
+## Overview
+
+Basic Memory allows you and users to record context in local Markdown files, building a rich knowledge base through natural conversations. The system automatically creates a semantic knowledge graph from simple text patterns.
+
+- **Local-First**: All data is stored in plain text files on the user's computer
+- **Real-Time**: Users see content updates immediately
+- **Bi-Directional**: Both you and users can read and edit notes
+- **Semantic**: Simple patterns create a structured knowledge graph
+- **Persistent**: Knowledge persists across sessions and conversations
+
+## The Importance of the Knowledge Graph
+
+Basic Memory's value comes from connections between notes, not just the notes themselves. When writing notes, your primary goal should be creating a rich, interconnected knowledge graph.
+
+When creating content, focus on:
+
+1. **Increasing Semantic Density**: Add multiple observations and relations to each note
+2. **Using Accurate References**: Aim to reference existing entities by their exact titles
+3. **Creating Forward References**: Feel free to reference entities that don't exist yet - Basic Memory will resolve these when they're created later
+4. **Creating Bidirectional Links**: When appropriate, connect entities from both directions
+5. **Using Meaningful Categories**: Add semantic context with appropriate observation categories
+6. **Choosing Precise Relations**: Use specific relation types that convey meaning
+
+Remember that a knowledge graph with 10 heavily connected notes is more valuable than 20 isolated notes. Your job is to help build these connections.
+
+## Core Tools Reference
+
+### Essential Content Management
+
+**Writing knowledge** (most important tool):
+```
+write_note(
+    title="Search Design",
+    content="# Search Design\n...",
+    folder="specs",                     # Optional
+    tags=["search", "design"],          # v0.13.0: now searchable!
+    project="work-notes"                # v0.13.0: target specific project
+)
+```
+
+**Reading knowledge:**
+```
+read_note("Search Design")              # By title
+read_note("specs/search-design")        # By path
+read_note("memory://specs/search")      # By memory URL
+```
+
+**Viewing notes as formatted artifacts (Claude Desktop):**
+```
+view_note("Search Design")              # Creates readable artifact
+view_note("specs/search-design")        # By permalink
+view_note("memory://specs/search")      # By memory URL
+```
+
+**Incremental editing** (v0.13.0):
+```
+edit_note(
+    identifier="Search Design",         # Must be EXACT title/permalink (strict matching)
+    operation="append",                 # append, prepend, find_replace, replace_section
+    content="\n## New Section\nContent here..."
+)
+```
+**⚠️ Important:** `edit_note` requires exact identifiers (no fuzzy matching). Use `search_notes()` first if uncertain.
+
+**File organization** (v0.13.0):
+```
+move_note(
+    identifier="Old Note",              # Must be EXACT title/permalink (strict matching)
+    destination="archive/old-note.md"   # Folders created automatically
+)
+```
+**⚠️ Important:** `move_note` requires exact identifiers (no fuzzy matching). Use `search_notes()` first if uncertain.
+
+### Project Management (v0.13.0)
+
+```
+list_projects()                         # Show available projects
+switch_project("work-notes")            # Change active project
+get_current_project()                   # Current project info
+```
+
+### Search & Discovery
+
+```
+search_notes("authentication system")   # v0.13.0: includes frontmatter tags
+build_context("memory://specs/search")  # Follow knowledge graph connections
+recent_activity(timeframe="1 week")     # Check what's been updated
+```
+
+## memory:// URLs Explained
+
+Basic Memory uses a special URL format to reference entities in the knowledge graph:
+
+- `memory://title` - Reference by title
+- `memory://folder/title` - Reference by folder and title 
+- `memory://permalink` - Reference by permalink
+- `memory://path/relation_type/*` - Follow all relations of a specific type
+- `memory://path/*/target` - Find all entities with relations to target
+
+## Semantic Markdown Format
+
+Knowledge is encoded in standard markdown using simple patterns:
+
+**Observations** - Facts about an entity:
+```markdown
+- [category] This is an observation #tag1 #tag2 (optional context)
+```
+
+**Relations** - Links between entities:
+```markdown
+- relation_type [[Target Entity]] (optional context)
+```
+
+**Common Categories & Relation Types:**
+- Categories: `[idea]`, `[decision]`, `[question]`, `[fact]`, `[requirement]`, `[technique]`, `[recipe]`, `[preference]`
+- Relations: `relates_to`, `implements`, `requires`, `extends`, `part_of`, `pairs_with`, `inspired_by`, `originated_from`
+
+## When to Record Context
+
+**Always consider recording context when**:
+
+1. Users make decisions or reach conclusions
+2. Important information emerges during conversation
+3. Multiple related topics are discussed
+4. The conversation contains information that might be useful later
+5. Plans, tasks, or action items are mentioned
+
+**Protocol for recording context**:
+
+1. Identify valuable information in the conversation
+2. Ask the user: "Would you like me to record our discussion about [topic] in Basic Memory?"
+3. If they agree, use `write_note` to capture the information
+4. If they decline, continue without recording
+5. Let the user know when information has been recorded: "I've saved our discussion about [topic] to Basic Memory."
+
+## Understanding User Interactions
+
+Users will interact with Basic Memory in patterns like:
+
+1. **Creating knowledge**:
+   ```
+   Human: "Let's write up what we discussed about search."
+   
+   You: I'll create a note capturing our discussion about the search functionality.
+   [Use write_note() to record the conversation details]
+   ```
+
+2. **Referencing existing knowledge**:
+   ```
+   Human: "Take a look at memory://specs/search"
+   
+   You: I'll examine that information.
+   [Use build_context() to gather related information]
+   [Then read_note() to access specific content]
+   ```
+
+3. **Finding information**:
+   ```
+   Human: "What were our decisions about auth?"
+   
+   You: Let me find that information for you.
+   [Use search_notes() to find relevant notes]
+   [Then build_context() to understand connections]
+   ```
+
+4. **Editing existing notes (v0.13.0)**:
+   ```
+   Human: "Add a section about deployment to my API documentation"
+   
+   You: I'll add that section to your existing documentation.
+   [Use edit_note() with operation="append" to add new content]
+   ```
+
+5. **Project management (v0.13.0)**:
+   ```
+   Human: "Switch to my work project and show recent activity"
+   
+   You: I'll switch to your work project and check what's been updated recently.
+   [Use switch_project() then recent_activity()]
+   ```
+
+6. **File organization (v0.13.0)**:
+   ```
+   Human: "Move my old meeting notes to the archive folder"
+   
+   You: I'll organize those notes for you.
+   [Use move_note() to relocate files with database consistency]
+   ```
+
+## Key Things to Remember
+
+1. **Files are Truth**
+   - All knowledge lives in local files on the user's computer
+   - Users can edit files outside your interaction
+   - Changes need to be synced by the user (usually automatic)
+   - Always verify information is current with `recent_activity()`
+
+2. **Building Context Effectively**
+   - Start with specific entities
+   - Follow meaningful relations
+   - Check recent changes
+   - Build context incrementally
+   - Combine related information
+
+3. **Writing Knowledge Wisely**
+   - Same title+folder overwrites existing notes
+   - Structure with clear headings and semantic markup
+   - Use tags for searchability (v0.13.0: frontmatter tags indexed)
+   - Keep files organized in logical folders
+
+4. **Leverage v0.13.0 Features**
+   - **Edit incrementally**: Use `edit_note()` for small changes vs rewriting
+   - **Switch projects**: Change context when user mentions different work areas
+   - **Organize proactively**: Move old content to archive folders
+   - **Cross-project operations**: Create notes in specific projects while maintaining context
+
+## Common Knowledge Patterns
+
+### Capturing Decisions
+
+```markdown
+---
+title: Coffee Brewing Methods
+tags: [coffee, brewing, pour-over, techniques]  # v0.13.0: Now searchable!
+---
+
+# Coffee Brewing Methods
+
+## Context
+I've experimented with various brewing methods including French press, pour over, and espresso.
+
+## Decision
+Pour over is my preferred method for light to medium roasts because it highlights subtle flavors and offers more control over the extraction.
+
+## Observations
+- [technique] Blooming the coffee grounds for 30 seconds improves extraction #brewing
+- [preference] Water temperature between 195-205°F works best #temperature
+- [equipment] Gooseneck kettle provides better control of water flow #tools
+- [timing] Total brew time of 3-4 minutes produces optimal extraction #process
+
+## Relations
+- pairs_with [[Light Roast Beans]]
+- contrasts_with [[French Press Method]]
+- requires [[Proper Grinding Technique]]
+- part_of [[Morning Coffee Routine]]
+```
+
+### Recording Project Structure
+
+```markdown
+# Garden Planning
+
+## Overview
+This document outlines the garden layout and planting strategy for this season.
+
+## Observations
+- [structure] Raised beds in south corner for sun exposure #layout
+- [structure] Drip irrigation system installed for efficiency #watering
+- [pattern] Companion planting used to deter pests naturally #technique
+
+## Relations
+- contains [[Vegetable Section]]
+- contains [[Herb Garden]]
+- implements [[Organic Gardening Principles]]
+```
+
+### Technical Discussions
+
+```markdown
+# Recipe Improvement Discussion
+
+## Key Points
+Discussed strategies for improving the chocolate chip cookie recipe.
+
+## Observations
+- [issue] Cookies spread too thin when baked at 350°F #texture
+- [solution] Chilling dough for 24 hours improves flavor and reduces spreading #technique
+- [decision] Will use brown butter instead of regular butter #flavor
+
+## Relations
+- improves [[Basic Cookie Recipe]]
+- inspired_by [[Bakery-Style Cookies]]
+- pairs_with [[Homemade Ice Cream]]
+```
+
+## v0.13.0 Workflow Examples
+
+### Multi-Project Conversations
+
+**User:** "I need to update my work documentation and also add a personal recipe note."
+
+**Workflow:**
+1. `list_projects()` - Check available projects
+2. `write_note(title="Sprint Planning", project="work-notes")` - Work content
+3. `write_note(title="Weekend Recipes", project="personal")` - Personal content
+
+### Incremental Note Building
+
+**User:** "Add a troubleshooting section to my setup guide."
+
+**Workflow:**
+1. `edit_note(identifier="Setup Guide", operation="append", content="\n## Troubleshooting\n...")`
+
+**User:** "Update the authentication section in my API docs."
+
+**Workflow:**
+1. `edit_note(identifier="API Documentation", operation="replace_section", section="## Authentication")`
+
+### Smart File Organization
+
+**User:** "My notes are getting messy in the main folder."
+
+**Workflow:**
+1. `move_note("Old Meeting Notes", "archive/2024/old-meetings.md")`
+2. `move_note("Project Notes", "projects/client-work/notes.md")`
+
+### Creating Effective Relations
+
+When creating relations:
+1. **Reference existing entities** by their exact title: `[[Exact Title]]`
+2. **Create forward references** to entities that don't exist yet - they'll be linked automatically when created
+3. **Search first** to find existing entities to reference
+4. **Use meaningful relation types**: `implements`, `requires`, `part_of` vs generic `relates_to`
+
+**Example workflow:**
+1. `search_notes("travel")` to find existing travel-related notes
+2. Reference found entities: `- part_of [[Japan Travel Guide]]`
+3. Add forward references: `- located_in [[Tokyo]]` (even if Tokyo note doesn't exist yet)
+
+## Common Issues & Solutions
+
+**Missing Content:**
+- Try `search_notes()` with broader terms if `read_note()` fails
+- Use fuzzy matching: search for partial titles
+
+**Forward References:**
+- These are normal! Basic Memory links them automatically when target notes are created
+- Inform users: "I've created forward references that will be linked when you create those notes"
+
+**Sync Issues:**
+- If information seems outdated, suggest `basic-memory sync`
+- Use `recent_activity()` to check if content is current
+
+**Strict Mode for Edit/Move Operations:**
+- `edit_note()` and `move_note()` require **exact identifiers** (no fuzzy matching for safety)
+- If identifier not found: use `search_notes()` first to find the exact title/permalink
+- Error messages will guide you to find correct identifiers
+- Example workflow:
+  ```
+  # ❌ This might fail if identifier isn't exact
+  edit_note("Meeting Note", "append", "content")
+  
+  # ✅ Safe approach: search first, then use exact result
+  results = search_notes("meeting")
+  edit_note("Meeting Notes 2024", "append", "content")  # Use exact title from search
+  ```
+
+## Best Practices
+
+1. **Proactively Record Context**
+   - Offer to capture important discussions
+   - Record decisions, rationales, and conclusions
+   - Link to related topics
+   - Ask for permission first: "Would you like me to save our discussion about [topic]?"
+   - Confirm when complete: "I've saved our discussion to Basic Memory"
+
+2. **Create a Rich Semantic Graph**
+   - **Add meaningful observations**: Include at least 3-5 categorized observations in each note
+   - **Create deliberate relations**: Connect each note to at least 2-3 related entities
+   - **Use existing entities**: Before creating a new relation, search for existing entities
+   - **Verify wikilinks**: When referencing `[[Entity]]`, use exact titles of existing notes
+   - **Check accuracy**: Use `search_notes()` or `recent_activity()` to confirm entity titles
+   - **Use precise relation types**: Choose specific relation types that convey meaning (e.g., "implements" instead of "relates_to")
+   - **Consider bidirectional relations**: When appropriate, create inverse relations in both entities
+
+3. **Structure Content Thoughtfully**
+   - Use clear, descriptive titles
+   - Organize with logical sections (Context, Decision, Implementation, etc.)
+   - Include relevant context and background
+   - Add semantic observations with appropriate categories
+   - Use a consistent format for similar types of notes
+   - Balance detail with conciseness
+
+4. **Navigate Knowledge Effectively**
+   - Start with specific searches
+   - Follow relation paths
+   - Combine information from multiple sources
+   - Verify information is current
+   - Build a complete picture before responding
+
+5. **Help Users Maintain Their Knowledge**
+   - Suggest organizing related topics
+   - Identify potential duplicates
+   - Recommend adding relations between topics
+   - Offer to create summaries of scattered information
+   - Suggest potential missing relations: "I notice this might relate to [topic], would you like me to add that connection?"
+
+
+Built with ♥️ by Basic Machines

docs/ARCHITECTURE.md

@@ -1,442 +0,0 @@
-# Basic Memory Architecture
-
-This document describes the architectural patterns and composition structure of Basic Memory.
-
-## Overview
-
-Basic Memory is a local-first knowledge management system with three entrypoints:
-- **API** - FastAPI REST server for HTTP access
-- **MCP** - Model Context Protocol server for LLM integration
-- **CLI** - Typer command-line interface
-
-Each entrypoint uses a **composition root** pattern to manage configuration and dependencies.
-
-## Composition Roots
-
-### What is a Composition Root?
-
-A composition root is the single place in an application where dependencies are wired together. In Basic Memory, each entrypoint has its own composition root that:
-
-1. Reads configuration from `ConfigManager`
-2. Resolves runtime mode (local/test)
-3. Creates and provides dependencies to downstream code
-
-**Key principle**: Only composition roots read global configuration. All other modules receive configuration explicitly.
-
-### Container Structure
-
-Each entrypoint has a container dataclass in its package:
-
-```
-src/basic_memory/
-├── api/
-│   └── container.py      # ApiContainer
-├── mcp/
-│   └── container.py      # McpContainer
-├── cli/
-│   └── container.py      # CliContainer
-└── runtime.py            # RuntimeMode enum and resolver
-```
-
-### Container Pattern
-
-All containers follow the same structure:
-
-```python
-@dataclass
-class Container:
-    config: BasicMemoryConfig
-    mode: RuntimeMode
-
-    @classmethod
-    def create(cls) -> "Container":
-        """Create container by reading ConfigManager."""
-        config = ConfigManager().config
-        mode = resolve_runtime_mode(is_test_env=config.is_test_env)
-        return cls(config=config, mode=mode)
-
-    @property
-    def some_computed_property(self) -> bool:
-        """Derived values based on config and mode."""
-        return self.mode.is_local and self.config.some_setting
-
-# Module-level singleton
-_container: Container | None = None
-
-def get_container() -> Container:
-    if _container is None:
-        raise RuntimeError("Container not initialized")
-    return _container
-
-def set_container(container: Container) -> None:
-    global _container
-    _container = container
-```
-
-### Runtime Mode Resolution
-
-The `RuntimeMode` enum centralizes mode detection:
-
-```python
-class RuntimeMode(Enum):
-    LOCAL = "local"
-    CLOUD = "cloud"
-    TEST = "test"
-
-    @property
-    def is_cloud(self) -> bool:
-        return self == RuntimeMode.CLOUD
-
-    @property
-    def is_local(self) -> bool:
-        return self == RuntimeMode.LOCAL
-
-    @property
-    def is_test(self) -> bool:
-        return self == RuntimeMode.TEST
-```
-
-Resolution follows this precedence in local app flows: **TEST > LOCAL**
-
-```python
-def resolve_runtime_mode(is_test_env: bool) -> RuntimeMode:
-    if is_test_env:
-        return RuntimeMode.TEST
-    return RuntimeMode.LOCAL
-```
-
-**Note**: `RuntimeMode` determines global behavior (e.g., whether to start file sync).
-Per-project routing is orthogonal: individual projects can be set to `cloud` mode via `ProjectMode`,
-which affects client routing in `get_client(project_name=...)` without changing global runtime mode.
-`RuntimeMode.CLOUD` may remain for compatibility, but standard local runtime resolution does not select it.
-
-## Dependencies Package
-
-### Structure
-
-The `deps/` package provides FastAPI dependencies organized by feature:
-
-```
-src/basic_memory/deps/
-├── __init__.py       # Re-exports for backwards compatibility
-├── config.py         # Configuration access
-├── db.py             # Database/session management
-├── projects.py       # Project resolution
-├── repositories.py   # Data access layer
-├── services.py       # Business logic layer
-└── importers.py      # Import functionality
-```
-
-### Usage in Routers
-
-```python
-from basic_memory.deps.services import get_entity_service
-from basic_memory.deps.projects import get_project_config
-
-@router.get("/entities/{id}")
-async def get_entity(
-    id: int,
-    entity_service: EntityService = Depends(get_entity_service),
-    project: ProjectConfig = Depends(get_project_config),
-):
-    return await entity_service.get(id)
-```
-
-### Backwards Compatibility
-
-The old `deps.py` file still exists as a thin re-export shim:
-
-```python
-# deps.py - backwards compatibility shim
-from basic_memory.deps import *
-```
-
-New code should import from specific submodules (`basic_memory.deps.services`) for clarity.
-
-## MCP Tools Architecture
-
-### Typed API Clients
-
-MCP tools communicate with the API through typed clients that encapsulate HTTP paths and response validation:
-
-```
-src/basic_memory/mcp/clients/
-├── __init__.py       # Re-exports all clients
-├── base.py           # BaseClient with common logic
-├── knowledge.py      # KnowledgeClient - entity CRUD
-├── search.py         # SearchClient - search operations
-├── memory.py         # MemoryClient - context building
-├── directory.py      # DirectoryClient - directory listing
-├── resource.py       # ResourceClient - resource reading
-└── project.py        # ProjectClient - project management
-```
-
-### Client Pattern
-
-Each client encapsulates API paths and validates responses:
-
-```python
-class KnowledgeClient(BaseClient):
-    """Client for knowledge/entity operations."""
-
-    async def resolve_entity(self, identifier: str) -> int:
-        """Resolve identifier to entity ID."""
-        response = await call_get(
-            self.http_client,
-            f"{self._base_path}/resolve/{identifier}",
-        )
-        return int(response.text)
-
-    async def get_entity(self, entity_id: int) -> EntityResponse:
-        """Get entity by ID."""
-        response = await call_get(
-            self.http_client,
-            f"{self._base_path}/entities/{entity_id}",
-        )
-        return EntityResponse.model_validate(response.json())
-```
-
-### Tool → Client → API Flow
-
-```
-MCP Tool (thin adapter)
-    ↓
-Typed Client (encapsulates paths, validates responses)
-    ↓
-HTTP API (FastAPI router)
-    ↓
-Service Layer (business logic)
-    ↓
-Repository Layer (data access)
-```
-
-Example tool using typed client:
-
-```python
-@mcp.tool()
-async def search_notes(
-    query: str,
-    project: str | None = None,
-    metadata_filters: dict | None = None,
-    tags: list[str] | None = None,
-    status: str | None = None,
-) -> SearchResponse:
-    async with get_project_client(project, context) as (client, active_project):
-        # Import client inside function to avoid circular imports
-        from basic_memory.mcp.clients import SearchClient
-        from basic_memory.schemas.search import SearchQuery
-
-        search_query = SearchQuery(
-            text=query,
-            metadata_filters=metadata_filters,
-            tags=tags,
-            status=status,
-        )
-        search_client = SearchClient(client, active_project.external_id)
-        return await search_client.search(search_query.model_dump())
-```
-
-### Per-Project Client Routing
-
-`get_project_client()` from `mcp/project_context.py` is an async context manager that:
-1. Resolves the project name from config (no network call)
-2. Creates the correctly-routed client based on the project's mode (local ASGI or cloud HTTP with API key)
-3. Validates the project via the API
-4. Yields `(client, active_project)` tuple
-
-This solves the bootstrap problem: you need the project name to choose the right client (local vs cloud), but you need the client to validate the project exists.
-
-```python
-from basic_memory.mcp.project_context import get_project_client
-
-async with get_project_client(project, context) as (client, active_project):
-    # client is routed based on project's mode (local or cloud)
-    # active_project is validated via the API
-    ...
-```
-
-## Sync Coordination
-
-### SyncCoordinator
-
-The `SyncCoordinator` centralizes sync/watch lifecycle management:
-
-```python
-@dataclass
-class SyncCoordinator:
-    """Coordinates file sync and watch operations."""
-
-    status: SyncStatus = SyncStatus.NOT_STARTED
-    sync_task: asyncio.Task | None = None
-    watch_service: WatchService | None = None
-
-    async def start(self, ...):
-        """Start sync and watch operations."""
-
-    async def stop(self):
-        """Stop all sync operations gracefully."""
-
-    def get_status_info(self) -> dict:
-        """Get current sync status for observability."""
-```
-
-### Status Enum
-
-```python
-class SyncStatus(Enum):
-    NOT_STARTED = "not_started"
-    STARTING = "starting"
-    RUNNING = "running"
-    STOPPING = "stopping"
-    STOPPED = "stopped"
-    ERROR = "error"
-```
-
-## Project Resolution
-
-### ProjectResolver
-
-Unified project selection across all entrypoints:
-
-```python
-class ProjectResolver:
-    """Resolves which project to use based on context."""
-
-    def resolve(
-        self,
-        explicit_project: str | None = None,
-    ) -> ResolvedProject:
-        """Resolve project using three-tier hierarchy:
-        1. Explicit project parameter
-        2. Default project from config
-        3. Single available project
-        """
-```
-
-### Resolution Modes
-
-```python
-class ResolutionMode(Enum):
-    EXPLICIT = "explicit"           # User specified project
-    DEFAULT = "default"             # Using configured default
-    SINGLE_PROJECT = "single"       # Only one project exists
-    FALLBACK = "fallback"           # Using first available
-```
-
-## Testing Patterns
-
-### Container Testing
-
-Each container has corresponding tests:
-
-```
-tests/
-├── api/test_api_container.py
-├── mcp/test_mcp_container.py
-└── cli/test_cli_container.py
-```
-
-Tests verify:
-- Container creation from config
-- Runtime mode properties
-- Container accessor functions (get/set)
-
-### Mocking Typed Clients
-
-When testing MCP tools, mock at the client level:
-
-```python
-def test_search_notes(monkeypatch):
-    import basic_memory.mcp.clients as clients_mod
-
-    class MockSearchClient:
-        async def search(self, query):
-            return SearchResponse(results=[...])
-
-    monkeypatch.setattr(clients_mod, "SearchClient", MockSearchClient)
-```
-
-## Design Principles
-
-### 1. Explicit Dependencies
-
-Modules receive configuration explicitly rather than reading globals:
-
-```python
-# Good - explicit injection
-async def sync_files(config: BasicMemoryConfig):
-    ...
-
-# Avoid - hidden global access
-async def sync_files():
-    config = ConfigManager().config  # Hidden coupling
-```
-
-### 2. Single Responsibility
-
-Each layer has a clear responsibility:
-- **Containers**: Wire dependencies
-- **Clients**: Encapsulate HTTP communication
-- **Services**: Business logic
-- **Repositories**: Data access
-- **Tools/Routers**: Thin adapters
-
-### 3. Deferred Imports
-
-To avoid circular imports, typed clients are imported inside functions:
-
-```python
-async def my_tool():
-    async with get_client() as client:
-        # Import here to avoid circular dependency
-        from basic_memory.mcp.clients import KnowledgeClient
-
-        knowledge_client = KnowledgeClient(client, project_id)
-```
-
-### 4. Backwards Compatibility
-
-When refactoring, maintain backwards compatibility via shims:
-
-```python
-# Old module becomes a shim
-from basic_memory.new_location import *
-
-# Docstring explains migration path
-"""
-DEPRECATED: Import from basic_memory.new_location instead.
-This shim will be removed in a future version.
-"""
-```
-
-## File Organization
-
-```
-src/basic_memory/
-├── api/
-│   ├── container.py          # API composition root
-│   ├── routers/              # FastAPI routers
-│   └── ...
-├── mcp/
-│   ├── container.py          # MCP composition root
-│   ├── clients/              # Typed API clients
-│   ├── tools/                # MCP tool definitions
-│   └── server.py             # MCP server setup
-├── cli/
-│   ├── container.py          # CLI composition root
-│   ├── app.py                # Typer app
-│   └── commands/             # CLI command groups
-├── deps/
-│   ├── config.py             # Config dependencies
-│   ├── db.py                 # Database dependencies
-│   ├── projects.py           # Project dependencies
-│   ├── repositories.py       # Repository dependencies
-│   ├── services.py           # Service dependencies
-│   └── importers.py          # Importer dependencies
-├── sync/
-│   ├── coordinator.py        # SyncCoordinator
-│   └── ...
-├── runtime.py                # RuntimeMode resolution
-├── project_resolver.py       # Unified project selection
-└── config.py                 # Configuration management
-```

docs/Docker.md

@@ -15,7 +15,7 @@ Basic Memory provides pre-built Docker images on GitHub Container Registry that
      --name basic-memory-server \
      -p 8000:8000 \
      -v /path/to/your/obsidian-vault:/app/data:rw \
-     -v basic-memory-config:/app/.basic-memory:rw \
+     -v basic-memory-config:/root/.basic-memory:rw \
      ghcr.io/basicmachines-co/basic-memory:latest
    ```
 
@@ -30,7 +30,7 @@ Basic Memory provides pre-built Docker images on GitHub Container Registry that
          - "8000:8000"
        volumes:
          - /path/to/your/obsidian-vault:/app/data:rw
-         - basic-memory-config:/app/.basic-memory:rw
+         - basic-memory-config:/root/.basic-memory:rw
        environment:
          - BASIC_MEMORY_DEFAULT_PROJECT=main
        restart: unless-stopped
@@ -67,7 +67,7 @@ docker build -t basic-memory .
 docker run -d \
   --name basic-memory-server \
   -v /path/to/your/obsidian-vault:/app/data:rw \
-  -v basic-memory-config:/app/.basic-memory:rw \
+  -v basic-memory-config:/root/.basic-memory:rw \
   -e BASIC_MEMORY_DEFAULT_PROJECT=main \
   basic-memory
 ```
@@ -86,11 +86,11 @@ Basic Memory requires several volume mounts for proper operation:
 
 2. **Configuration and Database** (Recommended):
    ```yaml
-   - basic-memory-config:/app/.basic-memory:rw
+   - basic-memory-config:/root/.basic-memory:rw
    ```
    Persistent storage for configuration and SQLite database.
 
-You can edit the basic-memory config.json file located in the /app/.basic-memory/config.json after Basic Memory starts.
+You can edit the basic-memory config.json file located in the /root/.basic-memory/config.json after Basic Memory starts.
 
 3. **Multiple Projects** (Optional):
    ```yaml
@@ -98,7 +98,7 @@ You can edit the basic-memory config.json file located in the /app/.basic-memory
    - /path/to/project2:/app/data/project2:rw
    ```
 
-You can edit the basic-memory config.json file located in the /app/.basic-memory/config.json
+You can edit the basic-memory config.json file located in the /root/.basic-memory/config.json
 
 ## CLI Commands via Docker
 
@@ -123,7 +123,7 @@ When using Docker volumes, you'll need to configure projects to point to your mo
 
 1. **Check current configuration:**
    ```bash
-   docker exec basic-memory-server cat /app/.basic-memory/config.json
+   docker exec basic-memory-server cat /root/.basic-memory/config.json
    ```
 
 2. **Add a project for your mounted volume:**
@@ -184,47 +184,16 @@ environment:
 
 ### Linux/macOS
 
-The Docker container now runs as a non-root user to avoid file ownership issues. By default, the container uses UID/GID 1000, but you can customize this to match your user:
+Ensure your knowledge directories have proper permissions:
 
 ```bash
-# Build with custom UID/GID to match your user
-docker build --build-arg UID=$(id -u) --build-arg GID=$(id -g) -t basic-memory .
+# Make directories readable/writable
+chmod -R 755 /path/to/your/obsidian-vault
 
-# Or use docker-compose with build args
+# If using specific user/group
+chown -R $USER:$USER /path/to/your/obsidian-vault
 ```
 
-**Example docker-compose.yml with custom user:**
-```yaml
-version: '3.8'
-services:
-  basic-memory:
-    build:
-      context: .
-      dockerfile: Dockerfile
-      args:
-        UID: 1000  # Replace with your UID
-        GID: 1000  # Replace with your GID
-    container_name: basic-memory-server
-    ports:
-      - "8000:8000"
-    volumes:
-      - /path/to/your/obsidian-vault:/app/data:rw
-      - basic-memory-config:/app/.basic-memory:rw
-    environment:
-      - BASIC_MEMORY_DEFAULT_PROJECT=main
-    restart: unless-stopped
-```
-
-**Using pre-built images:**
-If using the pre-built image from GitHub Container Registry, files will be created with UID/GID 1000. You can either:
-
-1. Change your local directory ownership to match:
-   ```bash
-   sudo chown -R 1000:1000 /path/to/your/obsidian-vault
-   ```
-
-2. Or build your own image with custom UID/GID as shown above.
-
 ### Windows
 
 When using Docker Desktop on Windows, ensure the directories are shared:
@@ -248,7 +217,7 @@ When using Docker Desktop on Windows, ensure the directories are shared:
       ```
 
 2. **Configuration Not Persisting:**
-    - Use named volumes for `/app/.basic-memory`
+    - Use named volumes for `/root/.basic-memory`
     - Check volume mount permissions
 
 3. **Network Connectivity:**
@@ -274,10 +243,10 @@ docker-compose logs -f basic-memory
 ## Security Considerations
 
 1. **Docker Security:**
-   The container runs as a non-root user (UID/GID 1000 by default) for improved security. You can customize the user ID using build arguments to match your local user.
+   The container runs as root for simplicity. For production, consider additional security measures.
 
 2. **Volume Permissions:**
-   Ensure mounted directories have appropriate permissions and don't expose sensitive data. With the non-root container, files will be created with the specified user ownership.
+   Ensure mounted directories have appropriate permissions and don't expose sensitive data.
 
 3. **Network Security:**
    If using HTTP transport, consider using reverse proxy with SSL/TLS and authentication if the endpoint is available on
@@ -319,7 +288,7 @@ For Docker-specific issues:
 1. Check the [troubleshooting section](#troubleshooting) above
 2. Review container logs: `docker-compose logs basic-memory`
 3. Verify volume mounts: `docker inspect basic-memory-server`
-4. Test file permissions: `docker exec basic-memory-server ls -la /app`
+4. Test file permissions: `docker exec basic-memory-server ls -la /root`
 
 For general Basic Memory support, see the main [README](../README.md)
 and [documentation](https://memory.basicmachines.co/).

docs/NOTE-FORMAT.md

@@ -1,494 +0,0 @@
-# Note Format Reference
-
-Every document in Basic Memory is a plain Markdown file. Files are the source of truth — changes to files automatically update the knowledge graph in the database. You maintain complete ownership, files work with git, and knowledge persists independently of any AI conversation.
-
-## Document Structure
-
-A note has three parts: YAML frontmatter, content (observations), and relations.
-
-```markdown
----
-title: Coffee Brewing Methods
-type: note
-tags: [coffee, brewing]
-permalink: coffee-brewing-methods
----
-
-# Coffee Brewing Methods
-
-## Observations
-- [method] Pour over provides more flavor clarity than French press
-- [technique] Water temperature at 205°F extracts optimal compounds #brewing
-- [preference] Ethiopian beans work well with lighter roasts (personal experience)
-
-## Relations
-- relates_to [[Coffee Bean Origins]]
-- requires [[Proper Grinding Technique]]
-- contrasts_with [[Tea Brewing Methods]]
-```
-
-The `## Observations` and `## Relations` headings are conventional but not required — the parser detects observations and relations by their syntax patterns anywhere in the document.
-
-## Frontmatter
-
-YAML metadata between `---` fences at the top of the file.
-
-| Field | Required | Default | Description |
-|-------|----------|---------|-------------|
-| `title` | No | filename stem | Used for linking and references. Auto-set from filename if missing. |
-| `type` | No | `note` | Entity type. Used for schema resolution and filtering. |
-| `tags` | No | `[]` | List or comma-separated string. Used for organization and search. |
-| `permalink` | No | generated from title | Stable identifier. Persists even if the file moves. |
-| `schema` | No | none | Schema attachment — dict (inline), string (reference), or omitted (implicit). |
-
-Custom fields are allowed. Any key not in the standard set is stored as `entity_metadata` and indexed for search and filtering.
-
-```yaml
----
-title: Paul Graham
-type: Person
-tags: [startups, essays, lisp]
-permalink: paul-graham
-status: active
-source: wikipedia
----
-```
-
-Here `status` and `source` are custom fields stored in `entity_metadata`.
-
-### Frontmatter Value Handling
-
-YAML automatically converts some values to native types. Basic Memory normalizes them:
-
-- Date strings (`2025-10-24`) → kept as ISO format strings
-- Numbers (`1.0`) → converted to strings
-- Booleans (`true`) → converted to strings (`"True"`)
-- Lists and dicts → preserved, items normalized recursively
-
-This prevents errors when downstream code expects string values.
-
-## Observations
-
-An observation is a categorized fact about the entity. Written as a Markdown list item.
-
-**Syntax:**
-
-```
-- [category] content text #tag1 #tag2 (context)
-```
-
-| Part | Required | Description |
-|------|----------|-------------|
-| `[category]` | Yes | Classification in square brackets. Any text except `[]()` chars. |
-| content | Yes | The fact or statement. |
-| `#tags` | No | Inline tags. Space-separated, each starting with `#`. |
-| `(context)` | No | Parenthesized text at end of line. Supporting details or source. |
-
-### Examples
-
-```markdown
-- [tech] Uses SQLite for storage #database
-- [design] Follows local-first architecture #architecture
-- [decision] Selected bcrypt for passwords #security (based on OWASP audit)
-- [name] Paul Graham
-- [expertise] Startups
-- [expertise] Lisp
-- [expertise] Essay writing
-```
-
-Array-like fields use repeated categories — multiple `[expertise]` observations above.
-
-### What Is Not an Observation
-
-The parser excludes these list item patterns:
-
-| Pattern | Example | Reason |
-|---------|---------|--------|
-| Checkboxes | `- [ ] Todo item`, `- [x] Done`, `- [-] Cancelled` | Task list syntax |
-| Markdown links | `- [text](url)` | URL link syntax |
-| Bare wiki links | `- [[Target]]` | Treated as a relation instead |
-
-A list item with `#tags` but no `[category]` is still parsed — the tags are extracted and the category defaults to `Note`.
-
-## Relations
-
-Relations connect documents to form the knowledge graph. There are two kinds.
-
-### Explicit Relations
-
-Written as list items with a relation type and a `[[wiki link]]` target.
-
-**Syntax:**
-
-```
-- relation_type [[Target Entity]] (context)
-```
-
-| Part | Required | Description |
-|------|----------|-------------|
-| `relation_type` | No | Text before `[[`. Defaults to `relates_to` if omitted. |
-| `[[Target]]` | Yes | Wiki link to the target entity. Matched by title or permalink. |
-| `(context)` | No | Parenthesized text after `]]`. Supporting details. |
-
-### Examples
-
-```markdown
-- implements [[Search Design]]
-- depends_on [[Database Schema]]
-- works_at [[Y Combinator]] (co-founder)
-- [[Some Entity]]
-```
-
-The last example — a bare `[[wiki link]]` in a list item — gets relation type `relates_to`.
-
-Common relation types:
-- `implements`, `depends_on`, `relates_to`, `inspired_by`
-- `extends`, `part_of`, `contains`, `pairs_with`
-- `works_at`, `authored`, `collaborated_with`
-
-Any text works as a relation type. These are conventions, not a fixed set.
-
-### Inline References
-
-Wiki links appearing in regular prose (not as list items) create implicit `links_to` relations.
-
-```markdown
-This builds on [[Core Design]] and uses [[Utility Functions]].
-```
-
-This creates two relations: `links_to [[Core Design]]` and `links_to [[Utility Functions]]`.
-
-### Forward References
-
-Relations can link to entities that don't exist yet. Basic Memory resolves them when the target is created.
-
-## Permalinks and memory:// URLs
-
-Every document has a unique **permalink** — a stable identifier derived from its title. You can set one explicitly in frontmatter, or let the system generate it.
-
-```yaml
-permalink: auth-approaches-2024
-```
-
-Permalinks form the basis of `memory://` URLs:
-
-```
-memory://auth-approaches-2024        # By permalink
-memory://Authentication Approaches   # By title (auto-resolves)
-memory://project/auth-approaches     # By path
-```
-
-Pattern matching is supported:
-
-```
-memory://auth*                       # Starts with "auth"
-memory://*/approaches                # Ends with "approaches"
-memory://project/*/requirements      # Nested wildcard
-```
-
-## Schemas
-
-Schemas declare the expected structure of a note — which observation categories and relation types a well-formed note should have. They use Picoschema, a compact notation from Google's Dotprompt that fits naturally in YAML frontmatter.
-
-### Picoschema Syntax
-
-```yaml
-schema:
-  name: string, full name              # required field with description
-  email?: string, contact email        # ? = optional
-  role?: string, job title
-  works_at?: Organization, employer    # capitalized type = entity reference
-  tags?(array): string, categories     # array of type
-  status?(enum): [active, inactive]    # enum with allowed values
-  metadata?(object):                   # nested object
-    updated_at?: string
-    source?: string
-```
-
-| Notation | Meaning | Example |
-|----------|---------|---------|
-| `field: type` | Required field | `name: string` |
-| `field?: type` | Optional field | `role?: string` |
-| `field(array): type` | Array of values | `expertise(array): string` |
-| `field?(enum): [vals]` | Enum with allowed values | `status?(enum): [active, inactive]` |
-| `field?(object):` | Nested object with sub-fields | `metadata?(object):` |
-| `, description` | Description after comma | `name: string, full name` |
-| `EntityName` | Capitalized type = entity reference | `works_at?: Organization` |
-
-**Scalar types:** `string`, `integer`, `number`, `boolean`, `any`
-
-Any type not in that set whose first letter is uppercase is treated as an entity reference (a relation target).
-
-### Schema-to-Note Mapping
-
-Schemas validate against existing observation/relation syntax. Note authors don't learn new syntax.
-
-| Schema Declaration | Maps To | Example in Note |
-|--------------------|---------|-----------------|
-| `field: string` | Observation `[field] value` | `- [name] Paul Graham` |
-| `field?(array): string` | Multiple `[field]` observations | `- [expertise] Lisp` (repeated) |
-| `field?: EntityType` | Relation `field [[Target]]` | `- works_at [[Y Combinator]]` |
-| `field?(array): EntityType` | Multiple `field` relations | `- authored [[Book]]` (repeated) |
-| `tags` | Frontmatter `tags` array | `tags: [startups, essays]` |
-| `field?(enum): [vals]` | Observation `[field] value` where value is in the set | `- [status] active` |
-
-Observations and relations not covered by the schema are valid — schemas describe a subset, not a straitjacket.
-
-### Schema Attachment
-
-Three ways to attach a schema to a note, resolved in priority order:
-
-**1. Inline schema** — `schema` is a dict in frontmatter:
-
-```yaml
----
-title: Team Standup 2024-01-15
-type: meeting
-schema:
-  attendees(array): string, who was there
-  decisions(array): string, what was decided
-  action_items(array): string, follow-ups
-  blockers?(array): string, anything stuck
----
-```
-
-Good for one-off structured notes or prototyping a schema before extracting it.
-
-**2. Explicit reference** — `schema` is a string naming a schema note:
-
-```yaml
----
-title: Basic Memory
-schema: SoftwareProject
----
-```
-
-or by permalink:
-
-```yaml
----
-title: LLM Memory Patterns
-schema: schema/research-project
----
-```
-
-Use when the note's `type` differs from the schema it should validate against, or when multiple schema variants exist.
-
-**3. Implicit by type** — no `schema` field, resolved by matching `type`:
-
-```yaml
----
-title: Paul Graham
-type: Person
----
-```
-
-The system looks up a schema note where `entity: Person`. If found, it applies. If not, no validation occurs.
-
-**4. No schema** — perfectly fine. Most notes don't need one.
-
-### Schema Notes
-
-A schema is itself a Basic Memory note with `type: schema`. It lives anywhere (though `schema/` is the conventional directory).
-
-```yaml
-# schema/Person.md
----
-title: Person
-type: schema
-entity: Person
-version: 1
-schema:
-  name: string, full name
-  role?: string, job title or position
-  works_at?: Organization, employer
-  expertise?(array): string, areas of knowledge
-  email?: string, contact email
-settings:
-  validation: warn
----
-
-# Person
-
-A human individual in the knowledge graph.
-```
-
-| Field | Required | Description |
-|-------|----------|-------------|
-| `type` | Yes | Must be `schema` |
-| `entity` | Yes | The entity type this schema describes (e.g., `Person`) |
-| `version` | No | Schema version number (default: `1`) |
-| `schema` | Yes | Picoschema dict defining the fields |
-| `settings.validation` | No | Validation mode (default: `warn`) |
-
-Schema notes are regular notes — they show up in search, can have observations and relations, and participate in the knowledge graph.
-
-### Validation Modes
-
-| Mode | Behavior |
-|------|----------|
-| `warn` | Warnings in output, doesn't block (default) |
-| `strict` | Errors that block sync, for CI/CD enforcement |
-| `off` | No validation |
-
-### Validation Output
-
-```
-$ bm schema validate people/ada-lovelace.md
-
-⚠ Person schema validation:
-  - Missing required field: name (expected [name] observation)
-  - Missing optional field: role
-  - Missing optional field: works_at (no relation found)
-
-ℹ Unmatched observations: [fact] ×2, [born] ×1
-ℹ Unmatched relations: collaborated_with
-```
-
-"Unmatched" items are informational — observations and relations the schema doesn't cover.
-
-### Schema Inference
-
-Generate schemas from existing notes by analyzing observation and relation frequency:
-
-```
-$ bm schema infer Person
-
-Analyzing 30 notes with type: Person...
-
-Observations found:
-  [name]        30/30  100%  → name: string
-  [role]        27/30   90%  → role?: string
-  [expertise]   18/30   60%  → expertise?(array): string
-  [email]        8/30   27%  → email?: string
-
-Relations found:
-  works_at      22/30   73%  → works_at?: Organization
-
-Suggested schema:
-  name: string, full name
-  role?: string, job title
-  expertise?(array): string, areas of knowledge
-  email?: string, contact email
-  works_at?: Organization, employer
-
-Save to schema/Person.md? [y/n]
-```
-
-Frequency thresholds:
-- **100% present** → required field
-- **25%+ present** → optional field
-- **Below 25%** → excluded from suggestion
-
-### Schema Drift Detection
-
-Track how usage patterns shift over time:
-
-```
-$ bm schema diff Person
-
-Schema drift detected:
-
-+ expertise: now in 81% of notes (was 12%)
-- department: dropped to 3% of notes
-~ works_at: cardinality changed (one → many)
-
-Update schema? [y/n/review]
-```
-
-## Complete Examples
-
-### Simple Note (No Schema)
-
-```markdown
----
-title: Project Ideas
-type: note
-tags: [ideas, brainstorm]
----
-
-# Project Ideas
-
-## Observations
-- [idea] Build a CLI tool for markdown linting #tooling
-- [idea] Create a recipe knowledge base #cooking
-- [priority] Focus on developer tools first (Q1 goal)
-
-## Relations
-- inspired_by [[Developer Workflow Research]]
-- part_of [[Q1 Planning]]
-```
-
-### Schema-Validated Note
-
-Schema at `schema/Person.md`:
-
-```yaml
----
-title: Person
-type: schema
-entity: Person
-version: 1
-schema:
-  name: string, full name
-  role?: string, job title or position
-  works_at?: Organization, employer
-  expertise?(array): string, areas of knowledge
-  email?: string, contact email
-settings:
-  validation: warn
----
-
-# Person
-
-A human individual in the knowledge graph.
-```
-
-Note at `people/paul-graham.md`:
-
-```markdown
----
-title: Paul Graham
-type: Person
-tags: [startups, essays, lisp]
----
-
-# Paul Graham
-
-## Observations
-- [name] Paul Graham
-- [role] Essayist and investor
-- [expertise] Startups
-- [expertise] Lisp
-- [expertise] Essay writing
-- [fact] Created Viaweb, the first web app
-
-## Relations
-- works_at [[Y Combinator]]
-- authored [[Hackers and Painters]]
-```
-
-The `[fact]` observation and `authored` relation are not in the schema — they're valid, just unmatched. The schema only checks that `[name]` exists (required) and looks for optional fields like `[role]`, `[expertise]`, and `works_at`.
-
-### Inline Schema Note
-
-```markdown
----
-title: Team Standup 2024-01-15
-type: meeting
-schema:
-  attendees(array): string, who was there
-  decisions(array): string, what was decided
-  action_items(array): string, follow-ups
-  blockers?(array): string, anything stuck
----
-
-# Team Standup 2024-01-15
-
-## Observations
-- [attendees] Paul
-- [attendees] Sarah
-- [decisions] Ship v2 by Friday
-- [action_items] Paul to review PR #42
-- [blockers] Waiting on API credentials
-```

docs/SPEC-PER-PROJECT-ROUTING.md

@@ -1,147 +0,0 @@
-# Simplified Local/Cloud Routing
-
-## Context
-
-Basic Memory now uses explicit, project-aware routing without a global cloud-mode toggle.
-Routing is determined by command-level flags and project mode, not by a global `cloud_mode` state.
-
-This document is the canonical contract for local/cloud routing behavior in CLI, MCP, and API-adjacent clients.
-
-## Goals
-
-1. Remove global `cloud_mode` from runtime/routing semantics.
-2. Keep MCP HTTP/SSE local-only; let stdio honor per-project routing.
-3. Make CLI routing explicit and easy to reason about.
-4. Support projects that exist in both local and cloud without ambiguity.
-
-## Routing Contract
-
-Routing is resolved in this order:
-
-1. Injected client factory (for composition/integration contexts)
-2. Explicit routing override (`--local` / `--cloud` or env vars below)
-3. Project-scoped routing (`project.mode`) when a project is known
-4. Default local routing
-
-### Routing Environment Variables
-
-- `BASIC_MEMORY_FORCE_LOCAL=true`: force local transport
-- `BASIC_MEMORY_FORCE_CLOUD=true`: force cloud proxy transport
-- `BASIC_MEMORY_EXPLICIT_ROUTING=true`: marks routing as explicitly chosen for this command
-
-When explicit routing is active, project mode does not override the selected route.
-
-## Config Semantics
-
-- `project.mode` is the only config-based routing signal for project-scoped operations.
-- Legacy `cloud_mode` values may be encountered during migration/loading but are not used for routing behavior.
-- Normalization saves remove stale `cloud_mode` from `~/.basic-memory/config.json`.
-
-### Example Config
-
-```json
-{
-  "projects": {
-    "main": {
-      "path": "/Users/me/basic-memory",
-      "mode": "local",
-      "local_sync_path": null,
-      "bisync_initialized": false,
-      "last_sync": null
-    },
-    "specs": {
-      "path": "specs",
-      "mode": "cloud",
-      "local_sync_path": "/Users/me/dev/specs",
-      "bisync_initialized": true,
-      "last_sync": "2026-02-06T17:36:38.544153"
-    }
-  },
-  "default_project": "main",
-  "cloud_api_key": "bmc_abc123...",
-  "cloud_host": "https://cloud.basicmemory.com"
-}
-```
-
-## Cloud Commands Are Auth-Only
-
-`bm cloud login`, `bm cloud logout`, and `bm cloud status` manage authentication state.
-
-- `bm cloud login`
-  - performs OAuth device flow
-  - stores/refreshes token material
-  - may verify cloud health/subscription
-  - does not change routing defaults
-- `bm cloud logout`
-  - removes stored OAuth session tokens
-  - does not change routing defaults
-- `bm cloud status`
-  - reports auth state (API key, OAuth token validity)
-  - runs health checks only when credentials are available
-
-## MCP Transport Routing
-
-### Stdio (default)
-
-`bm mcp --transport stdio` uses natural per-project routing.
-
-- Local-mode projects route through the in-process ASGI transport.
-- Cloud-mode projects route to the cloud proxy with Bearer auth (API key).
-- No explicit routing env vars are injected by the CLI command.
-- Externally-set env vars are honored (e.g. `BASIC_MEMORY_FORCE_CLOUD=true` for cloud deployments).
-- Users who need all projects forced local can set `BASIC_MEMORY_FORCE_LOCAL=true` externally.
-
-### HTTP and SSE Transports
-
-`bm mcp --transport streamable-http` and `bm mcp --transport sse` always route locally.
-
-These transports set explicit local routing (`BASIC_MEMORY_FORCE_LOCAL=true` and
-`BASIC_MEMORY_EXPLICIT_ROUTING=true`) before starting the server. This prevents cloud
-routing regardless of project mode, since HTTP/SSE serve as local API endpoints.
-
-## Project List UX for Dual Presence
-
-Projects may exist in both local and cloud. `bm project list` should display that clearly in one row per logical
-project identity, with explicit source/target signals.
-
-Recommended display contract:
-
-1. Keep one row per normalized project name/permalink.
-2. Show both local and cloud presence as separate columns/indicators.
-3. Show an explicit `MCP (stdio)` target column that always resolves to `local`.
-4. Keep CLI route semantics explicit:
-   - no flags: default local for non-project commands
-   - `--cloud`: force cloud
-   - `--local`: force local
-
-## Project LS Targeting
-
-`bm project ls` should clearly identify which project instance is being listed.
-
-Targeting rules:
-
-1. No routing flags: list local project files.
-2. `--cloud`: list cloud project files.
-3. `--local`: list local project files (explicit override).
-4. Output should label the active target (`LOCAL` or `CLOUD`) in heading or status line.
-
-## Runtime Mode
-
-Runtime mode is no longer a cloud/local routing switch for local app flows.
-
-- `resolve_runtime_mode(is_test_env)` resolves to:
-  - `TEST` when running in test environment
-  - `LOCAL` otherwise
-- `RuntimeMode.CLOUD` may remain for compatibility with existing tests/call sites but is not selected by normal local
-  runtime resolution.
-
-## Verification Checklist
-
-1. Loading config with legacy `cloud_mode` succeeds.
-2. Saving config strips legacy `cloud_mode`.
-3. `--local/--cloud` always override per-project mode for that command.
-4. No-project + no-flags commands route local by default.
-5. `bm cloud login/logout` do not toggle routing behavior.
-6. `bm mcp` stdio routes per-project mode; HTTP/SSE remain local-forced.
-7. `bm project list` communicates dual local/cloud presence without ambiguity.
-8. `bm project ls` output identifies route target explicitly.

docs/character-handling.md

@@ -1,241 +0,0 @@
-# Character Handling and Conflict Resolution
-
-Basic Memory handles various character encoding scenarios and file naming conventions to provide consistent permalink generation and conflict resolution. This document explains how the system works and how to resolve common character-related issues.
-
-## Overview
-
-Basic Memory uses a sophisticated system to generate permalinks from file paths while maintaining consistency across different operating systems and character encodings. The system normalizes file paths and generates unique permalinks to prevent conflicts.
-
-## Character Normalization Rules
-
-### 1. Permalink Generation
-
-When Basic Memory processes a file path, it applies these normalization rules:
-
-```
-Original: "Finance/My Investment Strategy.md"
-Permalink: "finance/my-investment-strategy"
-```
-
-**Transformation process:**
-1. Remove file extension (`.md`)
-2. Convert to lowercase (case-insensitive)
-3. Replace spaces with hyphens
-4. Replace underscores with hyphens
-5. Handle international characters (transliteration for Latin, preservation for non-Latin)
-6. Convert camelCase to kebab-case
-
-### 2. International Character Support
-
-**Latin characters with diacritics** are transliterated:
-- `ø` → `o` (Søren → soren)
-- `ü` → `u` (Müller → muller)
-- `é` → `e` (Café → cafe)
-- `ñ` → `n` (Niño → nino)
-
-**Non-Latin characters** are preserved:
-- Chinese: `中文/测试文档.md` → `中文/测试文档`
-- Japanese: `日本語/文書.md` → `日本語/文書`
-
-## Common Conflict Scenarios
-
-### 1. Hyphen vs Space Conflicts
-
-**Problem:** Files with existing hyphens conflict with generated permalinks from spaces.
-
-**Example:**
-```
-File 1: "basic memory bug.md"     → permalink: "basic-memory-bug"
-File 2: "basic-memory-bug.md"    → permalink: "basic-memory-bug" (CONFLICT!)
-```
-
-**Resolution:** The system automatically resolves this by adding suffixes:
-```
-File 1: "basic memory bug.md"     → permalink: "basic-memory-bug"
-File 2: "basic-memory-bug.md"    → permalink: "basic-memory-bug-1"
-```
-
-**Best Practice:** Choose consistent naming conventions within your project.
-
-### 2. Case Sensitivity Conflicts
-
-**Problem:** Different case variations that normalize to the same permalink.
-
-**Example on macOS:**
-```
-Directory: Finance/investment.md
-Directory: finance/investment.md  (different on filesystem, same permalink)
-```
-
-**Resolution:** Basic Memory detects case conflicts and prevents them during sync operations with helpful error messages.
-
-**Best Practice:** Use consistent casing for directory and file names.
-
-### 3. Character Encoding Conflicts
-
-**Problem:** Different Unicode normalizations of the same logical character.
-
-**Example:**
-```
-File 1: "café.md" (é as single character)
-File 2: "café.md" (e + combining accent)
-```
-
-**Resolution:** Basic Memory normalizes Unicode characters using NFD normalization to detect these conflicts.
-
-### 4. Forward Slash Conflicts
-
-**Problem:** Forward slashes in frontmatter or file names interpreted as path separators.
-
-**Example:**
-```yaml
----
-permalink: finance/investment/strategy
----
-```
-
-**Resolution:** Basic Memory validates frontmatter permalinks and warns about path separator conflicts.
-
-## Error Messages and Troubleshooting
-
-### "UNIQUE constraint failed: entity.file_path, entity.project_id"
-
-**Cause:** Two entities trying to use the same file path within a project.
-
-**Common scenarios:**
-1. File move operation where destination is already occupied
-2. Case sensitivity differences on macOS
-3. Character encoding conflicts
-4. Concurrent file operations
-
-**Resolution steps:**
-1. Check for duplicate file names with different cases
-2. Look for files with similar names but different character encodings
-3. Rename conflicting files to have unique names
-4. Run sync again after resolving conflicts
-
-### "File path conflict detected during move"
-
-**Cause:** Enhanced conflict detection preventing potential database integrity violations.
-
-**What this means:** The system detected that moving a file would create a conflict before attempting the database operation.
-
-**Resolution:** Follow the specific guidance in the error message, which will indicate the type of conflict detected.
-
-## Best Practices
-
-### 1. File Naming Conventions
-
-**Recommended patterns:**
-- Use consistent casing (prefer lowercase)
-- Use hyphens instead of spaces for multi-word files
-- Avoid special characters that could conflict with path separators
-- Be consistent with directory structure casing
-
-**Examples:**
-```
-✅ Good:
-- finance/investment-strategy.md
-- projects/basic-memory-features.md
-- docs/api-reference.md
-
-❌ Problematic:
-- Finance/Investment Strategy.md  (mixed case, spaces)
-- finance/Investment Strategy.md  (inconsistent case)
-- docs/API/Reference.md          (mixed case directories)
-```
-
-### 2. Permalink Management
-
-**Custom permalinks in frontmatter:**
-```yaml
----
-type: knowledge
-permalink: custom-permalink-name
----
-```
-
-**Guidelines:**
-- Use lowercase permalinks
-- Use hyphens for word separation
-- Avoid path separators unless creating sub-paths
-- Ensure uniqueness within your project
-
-### 3. Directory Structure
-
-**Consistent casing:**
-```
-✅ Good:
-finance/
-  investment-strategies.md
-  portfolio-management.md
-
-❌ Problematic:  
-Finance/           (capital F)
-  investment-strategies.md
-finance/           (lowercase f) 
-  portfolio-management.md
-```
-
-## Migration and Cleanup
-
-### Identifying Conflicts
-
-Use Basic Memory's built-in conflict detection:
-
-```bash
-# Sync will report conflicts
-basic-memory sync
-
-# Check sync status for warnings
-basic-memory status
-```
-
-### Resolving Existing Conflicts
-
-1. **Identify conflicting files** from sync error messages
-2. **Choose consistent naming convention** for your project
-3. **Rename files** to follow the convention
-4. **Re-run sync** to verify resolution
-
-### Bulk Renaming Strategy
-
-For projects with many conflicts:
-
-1. **Backup your project** before making changes
-2. **Standardize on lowercase** file and directory names
-3. **Replace spaces with hyphens** in file names
-4. **Use consistent character encoding** (UTF-8)
-5. **Test sync after each batch** of changes
-
-## System Enhancements
-
-### Recent Improvements (v0.13+)
-
-1. **Enhanced conflict detection** before database operations
-2. **Improved error messages** with specific resolution guidance
-3. **Character normalization utilities** for consistent handling
-4. **File swap detection** for complex move scenarios
-5. **Proactive conflict warnings** during permalink resolution
-
-### Monitoring and Logging
-
-The system now provides detailed logging for conflict resolution:
-
-```
-DEBUG: Detected potential file path conflicts for 'Finance/Investment.md': ['finance/investment.md']
-WARNING: File path conflict detected during move: entity_id=123 trying to move from 'old.md' to 'new.md'
-```
-
-These logs help identify and resolve conflicts before they cause sync failures.
-
-## Support and Resources
-
-If you encounter character-related conflicts not covered in this guide:
-
-1. **Check the logs** for specific conflict details
-2. **Review error messages** for resolution guidance  
-3. **Report issues** with examples of the conflicting files
-4. **Consider the file naming best practices** outlined above
-
-The Basic Memory system is designed to handle most character conflicts automatically while providing clear guidance for manual resolution when needed.

docs/cloud-cli.md

@@ -1,855 +0,0 @@
-# Basic Memory Cloud CLI Guide
-
-The Basic Memory Cloud CLI provides seamless integration between local and cloud knowledge bases using **project-scoped synchronization**. Each project can optionally sync with the cloud, giving you fine-grained control over what syncs and where.
-
-## Overview
-
-The cloud CLI enables you to:
-- **Authenticate cloud access** - OAuth/API key credentials are stored locally for cloud operations
-- **Project-scoped sync** - Each project independently manages its sync configuration
-- **Explicit operations** - Sync only what you want, when you want
-- **Bidirectional sync** - Keep local and cloud in sync with rclone bisync
-- **Offline access** - Work locally, sync when ready
-
-## Prerequisites
-
-Before using Basic Memory Cloud, you need:
-
-- **Active Subscription**: An active Basic Memory Cloud subscription is required to access cloud features
-- **Subscribe**: Visit [https://basicmemory.com/subscribe](https://basicmemory.com/subscribe) to sign up
-- **Optional**: Cloud is optional. Local-first open-source usage continues without cloud.
-- **OSS Discount**: Use code `{{OSS_DISCOUNT_CODE}}` for 20% off for 3 months.
-
-If you attempt to log in without an active subscription, you'll receive a "Subscription Required" error with a link to subscribe.
-
-## Architecture: Project-Scoped Sync
-
-### The Problem
-
-**Old approach (SPEC-8):** All projects lived in a single `~/basic-memory-cloud-sync/` directory. This caused:
-- ❌ Directory conflicts between mount and bisync
-- ❌ Auto-discovery creating phantom projects
-- ❌ Confusion about what syncs and when
-- ❌ All-or-nothing sync (couldn't sync just one project)
-
-**New approach (SPEC-20):** Each project independently configures sync.
-
-### How It Works
-
-**Projects can exist in three states:**
-
-1. **Cloud-only** - Project exists on cloud, no local copy
-2. **Cloud + Local (synced)** - Project has a local working directory that syncs
-3. **Local-only** - Project exists locally and is not routed to cloud
-
-**Example:**
-
-```bash
-# You have 3 projects on cloud:
-# - research: wants local sync at ~/Documents/research
-# - work: wants local sync at ~/work-notes
-# - temp: cloud-only, no local sync needed
-
-bm project add research --cloud --local-path ~/Documents/research
-bm project add work --cloud --local-path ~/work-notes
-bm project add temp --cloud  # No local sync
-
-# Now you can sync individually (after initial --resync):
-bm project bisync --name research
-bm project bisync --name work
-# temp stays cloud-only
-```
-
-**What happens under the covers:**
-- Config stores `cloud_projects` dict mapping project names to local paths
-- Each project gets its own bisync state in `~/.basic-memory/bisync-state/{project}/`
-- Rclone syncs using single remote: `basic-memory-cloud`
-- Projects can live anywhere on your filesystem, not forced into sync directory
-
-## Quick Start
-
-### 1. Authenticate Cloud Access
-
-Authenticate with cloud:
-
-```bash
-bm cloud login
-```
-
-**What this does:**
-1. Opens browser to Basic Memory Cloud authentication page
-2. Stores authentication tokens in `~/.basic-memory/basic-memory-cloud.json`
-3. Validates your subscription status
-4. Leaves routing behavior unchanged (auth only)
-
-**Result:** Cloud credentials are available for cloud-routed commands.
-Apply OSS discount code `{{OSS_DISCOUNT_CODE}}` during checkout to receive 20% off for 3 months.
-
-### 2. Set Up Sync
-
-Install rclone and configure credentials:
-
-```bash
-bm cloud setup
-```
-
-**What this does:**
-1. Installs rclone automatically (if needed)
-2. Fetches your tenant information from cloud
-3. Generates scoped S3 credentials for sync
-4. Configures single rclone remote: `basic-memory-cloud`
-
-**Result:** You're ready to sync projects. No sync directories created yet - those come with project setup.
-
-### 3. Add Projects with Sync
-
-Create projects with optional local sync paths:
-
-```bash
-# Create cloud project without local sync
-bm project add research --cloud
-
-# Create cloud project WITH local sync
-bm project add research --cloud --local-path ~/Documents/research
-
-# Or configure sync for existing project
-bm project sync-setup research ~/Documents/research
-```
-
-**What happens under the covers:**
-
-When you add a project with `--local-path`:
-1. Project created on cloud at `/app/data/research`
-2. Local path stored in config for that project (`local_sync_path`)
-3. Local directory created if it doesn't exist
-4. Bisync state directory created at `~/.basic-memory/bisync-state/research/`
-
-**Result:** Project is ready to sync, but no files synced yet.
-
-### 4. Sync Your Project
-
-Establish the initial sync baseline. **Best practice:** Always preview with `--dry-run` first:
-
-```bash
-# Step 1: Preview the initial sync (recommended)
-bm project bisync --name research --resync --dry-run
-
-# Step 2: If all looks good, run the actual sync
-bm project bisync --name research --resync
-```
-
-**What happens under the covers:**
-1. Rclone reads from `~/Documents/research` (local)
-2. Connects to `basic-memory-cloud:bucket-name/app/data/research` (remote)
-3. Creates bisync state files in `~/.basic-memory/bisync-state/research/`
-4. Syncs files bidirectionally with settings:
-   - `conflict_resolve=newer` (most recent wins)
-   - `max_delete=25` (safety limit)
-   - Respects `.bmignore` patterns
-
-**Result:** Local and cloud are in sync. Baseline established.
-
-**Why `--resync`?** This is an rclone requirement for the first bisync run. It establishes the initial state that future syncs will compare against. After the first sync, never use `--resync` unless you need to force a new baseline.
-
-See: https://rclone.org/bisync/#resync
-```
---resync
-This will effectively make both Path1 and Path2 filesystems contain a matching superset of all files. By default, Path2 files that do not exist in Path1 will be copied to Path1, and the process will then copy the Path1 tree to Path2.
-```
-
-### 5. Subsequent Syncs
-
-After the first sync, just run bisync without `--resync`:
-
-```bash
-bm project bisync --name research
-```
-
-**What happens:**
-1. Rclone compares local and cloud states
-2. Syncs changes in both directions
-3. Auto-resolves conflicts (newer file wins)
-4. Updates `last_sync` timestamp in config
-
-**Result:** Changes flow both ways - edit locally or in cloud, both stay in sync.
-
-### 6. Verify Setup
-
-Check status:
-
-```bash
-bm cloud status
-```
-
-You should see:
-- `OAuth: token valid` (or missing/expired)
-- `API Key: configured` (or not set)
-- `Cloud instance is healthy`
-- Instructions for project sync commands
-
-## Working with Projects
-
-### Understanding Project Commands
-
-**Key concept:** Use regular `bm project` commands (not `bm cloud project`).
-
-```bash
-# Local route
-bm project list --local
-bm project add research ~/Documents/research
-
-# Cloud route
-bm project list --cloud
-bm project add research --cloud
-```
-
-### Creating Projects
-
-**Use case 1: Cloud-only project (no local sync)**
-
-```bash
-bm project add temp-notes --cloud
-```
-
-**What this does:**
-- Creates project on cloud at `/app/data/temp-notes`
-- No local directory created
-- No sync configuration
-
-**Result:** Project exists on cloud, accessible via MCP tools, but no local copy.
-
-**Use case 2: Cloud project with local sync**
-
-```bash
-bm project add research --cloud --local-path ~/Documents/research
-```
-
-**What this does:**
-- Creates project on cloud at `/app/data/research`
-- Creates local directory `~/Documents/research`
-- Stores sync config in `~/.basic-memory/config.json`
-- Prepares for bisync (but doesn't sync yet)
-
-**Result:** Project ready to sync. Run `bm project bisync --name research --resync` to establish baseline.
-
-**Use case 3: Add sync to existing cloud project**
-
-```bash
-# Project already exists on cloud
-bm project sync-setup research ~/Documents/research
-```
-
-**What this does:**
-- Updates existing project's sync configuration
-- Creates local directory
-- Prepares for bisync
-
-**Result:** Existing cloud project now has local sync path. Run bisync to pull files down.
-
-### Listing Projects
-
-View all projects:
-
-```bash
-bm project list
-```
-
-**What you see:**
-- Local projects always
-- Cloud projects when credentials are available
-- Default project marked
-- Route-related metadata (for example, local/cloud presence and sync info)
-
-Example shape (single row for dual-presence projects):
-
-```text
-Name   Path            Local Path           Cloud Path   CLI Default   MCP (stdio)
-main   /basic-memory   ~/basic-memory       /basic-memory   local       local
-specs  /specs          ~/dev/specs          /specs          cloud       local
-```
-
-### When a Project Exists in Both Local and Cloud
-
-Use routing flags to disambiguate command targets:
-
-```bash
-# Force local target for this command
-bm project info main --local
-bm project ls --name main --local
-
-# Force cloud target for this command
-bm project info main --cloud
-bm project ls --name main --cloud
-```
-
-Default behavior for no-project, no-flag commands is local.
-For MCP stdio, routing is always local.
-
-## File Synchronization
-
-### Understanding the Sync Commands
-
-**There are three sync-related commands:**
-
-1. `bm project sync` - One-way: local → cloud (make cloud match local)
-2. `bm project bisync` - Two-way: local ↔ cloud (recommended)
-3. `bm project check` - Verify files match (no changes)
-
-### One-Way Sync: Local → Cloud
-
-**Use case:** You made changes locally and want to push to cloud (overwrite cloud).
-
-```bash
-bm project sync --name research
-```
-
-**What happens:**
-1. Reads files from `~/Documents/research` (local)
-2. Uses rclone sync to make cloud identical to local
-3. Respects `.bmignore` patterns
-4. Shows progress bar
-
-**Result:** Cloud now matches local exactly. Any cloud-only changes are overwritten.
-
-**When to use:**
-- You know local is the source of truth
-- You want to force cloud to match local
-- You don't care about cloud changes
-
-### Two-Way Sync: Local ↔ Cloud (Recommended)
-
-**Use case:** You edit files both locally and in cloud UI, want both to stay in sync.
-
-```bash
-# First time - establish baseline
-bm project bisync --name research --resync
-
-# Subsequent syncs
-bm project bisync --name research
-```
-
-**What happens:**
-1. Compares local and cloud states using bisync metadata
-2. Syncs changes in both directions
-3. Auto-resolves conflicts (newer file wins)
-4. Detects excessive deletes and fails safely (max 25 files)
-
-**Conflict resolution example:**
-
-```bash
-# Edit locally
-echo "Local change" > ~/Documents/research/notes.md
-
-# Edit same file in cloud UI
-# Cloud now has: "Cloud change"
-
-# Run bisync
-bm project bisync --name research
-
-# Result: Newer file wins (based on modification time)
-# If cloud was more recent, cloud version kept
-# If local was more recent, local version kept
-```
-
-**When to use:**
-- Default workflow for most users
-- You edit in multiple places
-- You want automatic conflict resolution
-
-### Verify Sync Integrity
-
-**Use case:** Check if local and cloud match without making changes.
-
-```bash
-bm project check --name research
-```
-
-**What happens:**
-1. Compares file checksums between local and cloud
-2. Reports differences
-3. No files transferred
-
-**Result:** Shows which files differ. Run bisync to sync them.
-
-```bash
-# One-way check (faster)
-bm project check --name research --one-way
-```
-
-### Preview Changes (Dry Run)
-
-**Use case:** See what would change without actually syncing.
-
-```bash
-bm project bisync --name research --dry-run
-```
-
-**What happens:**
-1. Runs bisync logic
-2. Shows what would be transferred/deleted
-3. No actual changes made
-
-**Result:** Safe preview of sync operations.
-
-### Advanced: List Project Files by Route
-
-**Use case:** Inspect local or cloud project files explicitly.
-
-```bash
-# List local project files (default target when no route flag is given)
-bm project ls --name research
-bm project ls --name research --local
-
-# List cloud project files
-bm project ls --name research --cloud
-
-# List files in subdirectory
-bm project ls --name research --cloud --path subfolder
-```
-
-**What happens:**
-1. Resolves route from flags (or local default when no route is given)
-2. Lists files for the chosen project instance
-3. No files transferred
-
-**Result:** See file listing for the target route.
-
-## Multiple Projects
-
-### Syncing Multiple Projects
-
-**Use case:** You have several projects with local sync, want to sync all at once.
-
-```bash
-# Setup multiple projects
-bm project add research --cloud --local-path ~/Documents/research
-bm project add work --cloud --local-path ~/work-notes
-bm project add personal --cloud --local-path ~/personal
-
-# Establish baselines
-bm project bisync --name research --resync
-bm project bisync --name work --resync
-bm project bisync --name personal --resync
-
-# Daily workflow: sync everything
-bm project bisync --name research
-bm project bisync --name work
-bm project bisync --name personal
-```
-
-**Future:** `--all` flag will sync all configured projects:
-
-```bash
-bm project bisync --all  # Coming soon
-```
-
-### Mixed Usage
-
-**Use case:** Some projects sync, some stay cloud-only.
-
-```bash
-# Projects with sync
-bm project add research --cloud --local-path ~/Documents/research
-bm project add work --cloud --local-path ~/work
-
-# Cloud-only projects
-bm project add archive --cloud
-bm project add temp-notes --cloud
-
-# Sync only the configured ones
-bm project bisync --name research
-bm project bisync --name work
-
-# Archive and temp-notes stay cloud-only
-```
-
-**Result:** Fine-grained control over what syncs.
-
-## Per-Project Cloud Routing (API Key)
-
-Route individual projects through cloud using an API key. This lets you keep some projects local while others route through cloud.
-
-### Setting Up API Key Auth
-
-**Option A: Create a key in the web app, then save it locally:**
-
-```bash
-bm cloud set-key bmc_abc123...
-```
-
-**Option B: Create a key via CLI (requires OAuth login first):**
-
-```bash
-bm cloud login                     # One-time OAuth login
-bm cloud create-key "my-laptop"    # Creates key and saves it locally
-```
-
-The API key is account-level — it grants access to all your cloud projects. It's stored in `~/.basic-memory/config.json` as `cloud_api_key`.
-
-### Setting Project Modes
-
-```bash
-# Route a project through cloud
-bm project set-cloud research
-
-# Revert to local mode
-bm project set-local research
-
-# View project modes
-bm project list
-```
-
-**What happens:**
-- `set-cloud`: validates the API key exists, then sets the project mode to `cloud` in config
-- `set-local`: reverts the project to local mode (removes the mode entry from config)
-- MCP tools and CLI commands for that project will route to `cloud_host/proxy` with the API key as Bearer token
-
-### How It Works
-
-When an MCP tool or CLI command runs for a cloud-mode project:
-
-1. `get_client(project_name="research")` checks the project's mode in config
-2. If mode is `cloud`, creates an HTTP client pointed at `cloud_host/proxy` with `Authorization: Bearer bmc_...`
-3. If mode is `local` (default), uses the in-process ASGI transport as usual
-
-**Routing priority** (highest to lowest):
-1. Factory injection (cloud app, tests)
-2. Explicit route override (`--local` / `--cloud`)
-3. Per-project cloud mode (API key)
-4. Local ASGI transport (default)
-
-Route override environment variables:
-- `BASIC_MEMORY_FORCE_LOCAL=true`
-- `BASIC_MEMORY_FORCE_CLOUD=true`
-- `BASIC_MEMORY_EXPLICIT_ROUTING=true`
-
-No-project, no-flag CLI commands default to local routing.
-
-### Configuration Example
-
-```json
-{
-  "projects": {
-    "personal": "/Users/me/notes",
-    "research": "/Users/me/research"
-  },
-  "project_modes": {
-    "research": "cloud"
-  },
-  "cloud_api_key": "bmc_abc123...",
-  "cloud_host": "https://cloud.basicmemory.com",
-  "default_project": "personal"
-}
-```
-
-In this example, `personal` stays local and `research` routes through cloud. Projects not listed in `project_modes` default to local.
-
-### Sync Behavior
-
-Cloud-mode projects are automatically skipped during local file sync (background sync and file watching). Their files live on the cloud instance, not locally.
-
-## OAuth Logout
-
-```bash
-bm cloud logout
-```
-
-**What this does:**
-1. Removes stored OAuth token(s)
-2. Does not change per-project route configuration
-3. Does not change command routing defaults
-
-**Result:** OAuth session is cleared. API-key-based routing still works if `cloud_api_key` is configured.
-
-## Filter Configuration
-
-### Understanding .bmignore
-
-**The problem:** You don't want to sync everything (e.g., `.git`, `node_modules`, database files).
-
-**The solution:** `.bmignore` file with gitignore-style patterns.
-
-**Location:** `~/.basic-memory/.bmignore`
-
-**Default patterns:**
-
-```gitignore
-# Version control
-.git/**
-
-# Python
-__pycache__/**
-*.pyc
-.venv/**
-venv/**
-
-# Node.js
-node_modules/**
-
-# Basic Memory internals
-memory.db/**
-memory.db-shm/**
-memory.db-wal/**
-config.json/**
-watch-status.json/**
-.bmignore.rclone/**
-
-# OS files
-.DS_Store/**
-Thumbs.db/**
-
-# Environment files
-.env/**
-.env.local/**
-```
-
-**How it works:**
-1. On first sync, `.bmignore` created with defaults
-2. Patterns converted to rclone filter format (`.bmignore.rclone`)
-3. Rclone uses filters during sync
-4. Same patterns used by all projects
-
-**Customizing:**
-
-```bash
-# Edit patterns
-code ~/.basic-memory/.bmignore
-
-# Add custom patterns
-echo "*.tmp/**" >> ~/.basic-memory/.bmignore
-
-# Next sync uses updated patterns
-bm project bisync --name research
-```
-
-## Troubleshooting
-
-### Authentication Issues
-
-**Problem:** "Authentication failed" or "Invalid token"
-
-**Solution:** Re-authenticate:
-
-```bash
-bm cloud logout
-bm cloud login
-```
-
-### Subscription Issues
-
-**Problem:** "Subscription Required" error
-
-**Solution:**
-1. Visit subscribe URL shown in error
-2. Sign up for subscription
-3. Run `bm cloud login` again
-
-**Note:** Access is immediate when subscription becomes active.
-
-### Bisync Initialization
-
-**Problem:** "First bisync requires --resync"
-
-**Explanation:** Bisync needs a baseline state before it can sync changes.
-
-**Solution:**
-
-```bash
-bm project bisync --name research --resync
-```
-
-**What this does:**
-- Establishes initial sync state
-- Creates baseline in `~/.basic-memory/bisync-state/research/`
-- Syncs all files bidirectionally
-
-**Result:** Future syncs work without `--resync`.
-
-### Empty Directory Issues
-
-**Problem:** "Empty prior Path1 listing. Cannot sync to an empty directory"
-
-**Explanation:** Rclone bisync doesn't work well with completely empty directories. It needs at least one file to establish a baseline.
-
-**Solution:** Add at least one file before running `--resync`:
-
-```bash
-# Create a placeholder file
-echo "# Research Notes" > ~/Documents/research/README.md
-
-# Now run bisync
-bm project bisync --name research --resync
-```
-
-**Why this happens:** Bisync creates listing files that track the state of each side. When both directories are completely empty, these listing files are considered invalid by rclone.
-
-**Best practice:** Always have at least one file (like a README.md) in your project directory before setting up sync.
-
-### Bisync State Corruption
-
-**Problem:** Bisync fails with errors about corrupted state or listing files
-
-**Explanation:** Sometimes bisync state can become inconsistent (e.g., after mixing dry-run and actual runs, or after manual file operations).
-
-**Solution:** Clear bisync state and re-establish baseline:
-
-```bash
-# Clear bisync state
-bm project bisync-reset research
-
-# Re-establish baseline
-bm project bisync --name research --resync
-```
-
-**What this does:**
-- Removes all bisync metadata from `~/.basic-memory/bisync-state/research/`
-- Forces fresh baseline on next `--resync`
-- Safe operation (doesn't touch your files)
-
-**Note:** This command also runs automatically when you remove a project to clean up state directories.
-
-### Too Many Deletes
-
-**Problem:** "Error: max delete limit (25) exceeded"
-
-**Explanation:** Bisync detected you're about to delete more than 25 files. This is a safety check to prevent accidents.
-
-**Solution 1:** Review what you're deleting, then force resync:
-
-```bash
-# Check what would be deleted
-bm project bisync --name research --dry-run
-
-# If correct, establish new baseline
-bm project bisync --name research --resync
-```
-
-**Solution 2:** Use one-way sync if you know local is correct:
-
-```bash
-bm project sync --name research
-```
-
-### Project Not Configured for Sync
-
-**Problem:** "Project research has no local_sync_path configured"
-
-**Explanation:** Project exists on cloud but has no local sync path.
-
-**Solution:**
-
-```bash
-bm project sync-setup research ~/Documents/research
-bm project bisync --name research --resync
-```
-
-### Connection Issues
-
-**Problem:** "Cannot connect to cloud instance"
-
-**Solution:** Check status:
-
-```bash
-bm cloud status
-```
-
-If instance is down, wait a few minutes and retry.
-
-## Security
-
-- **Authentication**: OAuth 2.1 with PKCE flow
-- **Tokens**: Stored securely in `~/.basic-memory/basic-memory-cloud.json`
-- **Transport**: All data encrypted in transit (HTTPS)
-- **Credentials**: Scoped S3 credentials (read-write to your tenant only)
-- **Isolation**: Your data isolated from other tenants
-- **Ignore patterns**: Sensitive files automatically excluded via `.bmignore`
-
-## Command Reference
-
-### Cloud Authentication
-
-```bash
-bm cloud login              # Authenticate and store OAuth credentials
-bm cloud logout             # Remove stored OAuth credentials
-bm cloud status             # Check auth state and instance health
-bm cloud promo --off        # Disable CLI cloud promo notices
-```
-
-### API Key Management
-
-```bash
-bm cloud set-key <key>      # Save a cloud API key (bmc_ prefixed)
-bm cloud create-key <name>  # Create API key via cloud API (requires OAuth login)
-```
-
-### Setup
-
-```bash
-bm cloud setup              # Install rclone and configure credentials
-```
-
-### Project Management
-
-```bash
-bm project list --local                   # Local project list
-bm project list --cloud                   # Cloud project list
-bm project add <name> --cloud             # Create cloud project (no sync)
-bm project add <name> --cloud --local-path <path> # Create with local sync
-bm project sync-setup <name> <path>       # Add sync to existing project
-bm project rm <name>                      # Delete project
-```
-
-### Per-Project Routing
-
-```bash
-bm project set-cloud <name>  # Route project through cloud (requires API key)
-bm project set-local <name>  # Revert project to local mode
-```
-
-### File Synchronization
-
-```bash
-# One-way sync (local → cloud)
-bm project sync --name <project>
-bm project sync --name <project> --dry-run
-bm project sync --name <project> --verbose
-
-# Two-way sync (local ↔ cloud) - Recommended
-bm project bisync --name <project>          # After first --resync
-bm project bisync --name <project> --resync # First time / force baseline
-bm project bisync --name <project> --dry-run
-bm project bisync --name <project> --verbose
-
-# Integrity check
-bm project check --name <project>
-bm project check --name <project> --one-way
-
-# List project files by route
-bm project ls --name <project>          # Default target: local
-bm project ls --name <project> --local
-bm project ls --name <project> --cloud
-bm project ls --name <project> --cloud --path <subpath>
-```
-
-## Summary
-
-**Basic Memory Cloud uses project-scoped sync:**
-
-1. **Authenticate cloud access** - `bm cloud login`
-2. **Install rclone** - `bm cloud setup`
-3. **Add projects with sync** - `bm project add research --cloud --local-path ~/Documents/research`
-4. **Preview first sync** - `bm project bisync --name research --resync --dry-run`
-5. **Establish baseline** - `bm project bisync --name research --resync`
-6. **Daily workflow** - `bm project bisync --name research`
-
-**Key benefits:**
-- ✅ Each project independently syncs (or doesn't)
-- ✅ Projects can live anywhere on disk
-- ✅ Explicit sync operations (no magic)
-- ✅ Safe by design (max delete limits, conflict resolution)
-- ✅ Full offline access (work locally, sync when ready)
-
-**Future enhancements:**
-- `--all` flag to sync all configured projects
-- Project list showing sync status
-- Watch mode for automatic sync

docs/cloud-semantic-search-value.md

@@ -1,91 +0,0 @@
-# Cloud Semantic Search Value (Customer-Facing Technical Story)
-
-This document explains why teams should buy cloud semantic search even when local search exists.
-
-## Core Promise
-
-Markdown files remain the source of truth in both local and cloud modes.
-
-- Files are portable.
-- Search indexes are derived and rebuildable.
-- You never get locked into proprietary document storage.
-
-## The Customer Problem
-
-Teams paying for cloud are usually not optimizing for "can this run locally." They are optimizing for:
-
-- finding the right note the first time,
-- keeping retrieval quality high as note volume grows,
-- avoiding search slowdowns while content is actively changing,
-- getting consistent results across users, agents, and sessions.
-
-## Why Cloud Is the Aspirin
-
-Cloud semantic search is the immediate pain reliever because it fixes the problems users feel right now.
-
-### 1) Better hit rate on real queries
-
-Cloud uses stronger managed embeddings than the default local model, which improves semantic recall for paraphrases and vague questions.
-
-Customer outcome:
-
-- fewer "I know this exists but search missed it" moments,
-- less query rewording,
-- faster time to answer.
-
-### 2) Better behavior under active workloads
-
-Cloud indexing runs out of band in workers, so indexing does not compete with interactive read/write traffic.
-
-Customer outcome:
-
-- stable search responsiveness during heavy updates,
-- fresher semantic results shortly after edits,
-- less user-visible performance variance.
-
-### 3) Better consistency for shared knowledge
-
-Cloud retrieval runs against a centralized tenant index, so teams and agents resolve against the same semantic state.
-
-Customer outcome:
-
-- fewer "works on my machine" search differences,
-- more predictable agent behavior across environments,
-- easier cross-user collaboration on large knowledge bases.
-
-### 4) Better quality at higher scale
-
-With Postgres + `pgvector` per tenant, cloud can sustain larger note collections and higher query volumes than typical local setups.
-
-Customer outcome:
-
-- confidence as repositories grow to tens of thousands of notes,
-- less need for user-side tuning,
-- fewer quality regressions as usage increases.
-
-## Local Is the Vitamin
-
-Local semantic search still matters and should stay strong.
-
-- offline use,
-- privacy-first operation,
-- no cloud dependency,
-- user-controlled runtime.
-
-It compounds long-term ownership and resilience, but does not remove the immediate pain points cloud solves for teams at scale.
-
-## Recommended Messaging
-
-One-liner:
-
-"Cloud semantic search is the aspirin: it fixes retrieval quality and performance pain now. Local semantic search is the vitamin: it builds long-term control and resilience."
-
-Long form:
-
-"Basic Memory keeps markdown as the source of truth everywhere. Local gives privacy and offline control. Cloud adds immediate, measurable improvements in search quality, consistency, and responsiveness for teams and agents running at scale."
-
-## Packaging Guidance
-
-- Base: local FTS plus optional local semantic search.
-- Cloud value: higher semantic quality, stable performance under load, and consistent team-wide retrieval.
-- Keep interfaces pluggable (`EmbeddingProvider`, vector backend protocol) so implementation can evolve without changing user workflows.

docs/mcp-ui-bakeoff-instructions.md

@@ -1,138 +0,0 @@
-# MCP UI Bakeoff - Instructions & Test Plan
-
-Last updated: 2026-02-02
-
-## Scope
-
-Compare three presentation paths for Basic Memory MCP tools:
-
-1. **Tool‑UI (React)** via MCP App resources.
-2. **MCP‑UI Python SDK** embedded UI resources (legacy host path).
-3. **ASCII/ANSI** output for TUI clients.
-
-This doc is the running instruction set and test plan. Update as implementation progresses.
-
----
-
-## Prerequisites
-
-- Repo: `basic-memory` (worktree: `basic-memory-mcp-ui-poc`)
-- Node for tool‑ui build (already used for POC)
-- Python 3.12+ with `uv`
-
-Optional (for MCP‑UI Python SDK path):
-
-- Local repo: `/Users/phernandez/dev/mcp-ui`
-- Install the server SDK into the Basic Memory venv:
-  - `uv pip install -e /Users/phernandez/dev/mcp-ui/sdks/python/server`
-
----
-
-## Build / Refresh Steps
-
-### Tool‑UI React bundle
-
-```bash
-cd ui/tool-ui-react
-npm install
-npm run build
-```
-
-This regenerates:
-
-- `src/basic_memory/mcp/ui/html/search-results-tool-ui.html`
-- `src/basic_memory/mcp/ui/html/note-preview-tool-ui.html`
-
----
-
-## How to Run the MCP Server
-
-```bash
-basic-memory mcp --transport stdio
-```
-
-Optional to pick UI variant for MCP App resources:
-
-```bash
-export BASIC_MEMORY_MCP_UI_VARIANT=tool-ui   # or vanilla | mcp-ui
-```
-
----
-
-## Test Cases
-
-### 1) MCP App Resource UI (tool‑ui / vanilla / mcp‑ui)
-
-Tools:
-- `search_notes`
-- `read_note`
-
-Expect:
-- Tool meta points to `ui://basic-memory/search-results` and `ui://basic-memory/note-preview`
-- Resource content differs by `BASIC_MEMORY_MCP_UI_VARIANT`
-- Variant‑specific URIs also available:
-  - `ui://basic-memory/search-results/vanilla`
-  - `ui://basic-memory/search-results/tool-ui`
-  - `ui://basic-memory/search-results/mcp-ui`
-  - `ui://basic-memory/note-preview/vanilla`
-  - `ui://basic-memory/note-preview/tool-ui`
-  - `ui://basic-memory/note-preview/mcp-ui`
-
-Manual check:
-- Trigger tool in MCP‑App‑capable host and confirm UI renders.
-
----
-
-### 2) Text / JSON Output Modes
-
-Tools:
-- `search_notes(output_format="text" | "json")`
-- `read_note(output_format="text" | "json")`
-- `write_note(output_format="text" | "json")`
-- `edit_note(output_format="text" | "json")`
-- `recent_activity(output_format="text" | "json")`
-- `list_memory_projects(output_format="text" | "json")`
-- `create_memory_project(output_format="text" | "json")`
-- `delete_note(output_format="text" | "json")`
-- `move_note(output_format="text" | "json")`
-- `build_context(output_format="json" | "text")`
-
-Expect:
-- `text` mode preserves existing human-readable responses.
-- `json` mode returns structured dict/list payloads for machine-readable clients.
-
-Automated:
-- `uv run pytest test-int/mcp/test_output_format_json_integration.py`
-
----
-
-### 3) MCP‑UI Python SDK (embedded UI resource)
-
-Tools (embedded resource responses):
-- `search_notes_ui` (MCP‑UI SDK)
-- `read_note_ui` (MCP‑UI SDK)
-
-Expected output:
-- Tool response content contains an EmbeddedResource (`type: "resource"`)
-- `mimeType` is `text/html`
-- `_meta` includes:
-  - `mcpui.dev/ui-preferred-frame-size`
-  - `mcpui.dev/ui-initial-render-data`
-
-Manual check:
-- Render tool responses using `UIResourceRenderer` (legacy host flow).
-
-Automated (if SDK installed):
-- `uv run pytest test-int/mcp/test_ui_sdk_integration.py`
-
----
-
-## Bakeoff Notes Template
-
-Fill in after running:
-
-- Tool‑UI (React): __
-- MCP‑UI SDK (embedded): __
-- Text/JSON modes: __
-
-Decision + rationale: __

docs/metadata-search.md

@@ -1,260 +0,0 @@
-# Metadata Search Reference
-
-Basic Memory automatically indexes custom frontmatter fields so you can query them with structured filters. Any YAML key in a note's frontmatter beyond the standard set (`title`, `type`, `tags`, `permalink`, `schema`) is stored as `entity_metadata` and becomes searchable.
-
-## Querying with `search_notes`
-
-`search_notes` is the single search tool for all queries — text, metadata filters, or both. The `query` parameter is optional, so you can use metadata filters alone without passing an empty string.
-
-## Filter Syntax
-
-Filters are a JSON dictionary where each key targets a frontmatter field and the value specifies the match condition. Multiple keys combine with **AND** logic — every filter must match.
-
-### Equality
-
-Match a single value exactly.
-
-```json
-{"status": "active"}
-```
-
-Finds notes whose frontmatter contains `status: active`.
-
-### Array Contains (all)
-
-Pass a list to require **all** listed values to be present in the field.
-
-```json
-{"tags": ["security", "oauth"]}
-```
-
-Finds notes tagged with both `security` and `oauth`.
-
-### `$in` (any of)
-
-Match if the field equals **any** value in the list.
-
-```json
-{"priority": {"$in": ["high", "critical"]}}
-```
-
-### `$gt`, `$gte`, `$lt`, `$lte`
-
-Numeric and text comparisons. Numeric values use numeric comparison; strings use lexicographic comparison.
-
-```json
-{"confidence": {"$gt": 0.7}}
-{"score": {"$lte": 100}}
-```
-
-### `$between`
-
-Range filter (inclusive). Takes a `[min, max]` pair.
-
-```json
-{"score": {"$between": [0.3, 0.8]}}
-```
-
-### Nested Access (dot notation)
-
-Access nested frontmatter values using dots.
-
-```json
-{"schema.version": "2"}
-```
-
-This queries the `version` key inside a `schema` object in frontmatter.
-
-### Summary Table
-
-| Operator | Syntax | Example |
-|----------|--------|---------|
-| Equality | `{"field": "value"}` | `{"status": "active"}` |
-| Array contains (all) | `{"field": ["a", "b"]}` | `{"tags": ["security", "oauth"]}` |
-| `$in` (any of) | `{"field": {"$in": [...]}}` | `{"priority": {"$in": ["high", "critical"]}}` |
-| `$gt` / `$gte` | `{"field": {"$gt": N}}` | `{"confidence": {"$gt": 0.7}}` |
-| `$lt` / `$lte` | `{"field": {"$lt": N}}` | `{"score": {"$lt": 0.5}}` |
-| `$between` | `{"field": {"$between": [min, max]}}` | `{"score": {"$between": [0.3, 0.8]}}` |
-| Nested access | `{"a.b": "value"}` | `{"schema.version": "2"}` |
-
-**Key rules:**
-- Filter keys must match `[A-Za-z0-9_-]+` (dots separate nesting levels).
-- Each operator dict must contain exactly one operator.
-- `$in` and array-contains require non-empty lists.
-- `$between` requires exactly two values `[min, max]`.
-
-## MCP Tool — `search_notes`
-
-`search_notes` is the single search tool for text queries, metadata filters, or both. The `query` parameter is optional.
-
-**Relevant parameters:**
-
-| Parameter | Type | Description |
-|-----------|------|-------------|
-| `query` | string (optional) | Text search query. Omit for filter-only searches. |
-| `metadata_filters` | dict | Structured filter dict (see syntax above) |
-| `tags` | list[str] | Convenience shorthand — merged into `metadata_filters["tags"]` |
-| `status` | string | Convenience shorthand — merged into `metadata_filters["status"]` |
-
-**Merging rules:** `tags` and `status` are convenience shortcuts. They are merged into `metadata_filters` using `setdefault` — if the same key already exists in `metadata_filters`, the explicit filter wins.
-
-**Examples:**
-
-```python
-# Text search filtered by metadata
-await search_notes("authentication", metadata_filters={"status": "draft"})
-
-# Filter-only search (no query needed)
-await search_notes(metadata_filters={"type": "spec"})
-
-# Combine text, tags shortcut, and metadata
-await search_notes(
-    "oauth flow",
-    tags=["security"],
-    metadata_filters={"confidence": {"$gt": 0.7}},
-)
-
-# Convenience shortcuts
-await search_notes("planning", status="active")
-await search_notes(tags=["tier1", "alpha"])
-```
-
-## Tag Search Shortcuts
-
-The `tag:` prefix in a search query is a shorthand for tag-based metadata filtering. When `search_notes` receives a query starting with `tag:`, it converts the query into a `tags` filter and clears the text query.
-
-```python
-# These are equivalent:
-await search_notes("tag:tier1")
-await search_notes("", tags=["tier1"])
-
-# Multiple tags (comma or space separated) — all must be present:
-await search_notes("tag:tier1,alpha")
-await search_notes("tag:tier1 alpha")
-```
-
-## CLI Access
-
-The `bm tool search-notes` command exposes metadata filtering via `--meta` and `--filter` flags.
-
-### `--meta` — simple key=value filters
-
-Repeatable flag for equality filters on frontmatter fields.
-
-```bash
-# Single filter
-bm tool search-notes "my query" --meta status=draft
-
-# Multiple filters (AND logic)
-bm tool search-notes "" --meta status=active --meta priority=high
-```
-
-### `--filter` — advanced JSON filters
-
-Pass a full JSON filter dictionary for operator-based queries.
-
-```bash
-# Range filter
-bm tool search-notes "" --filter '{"score": {"$between": [0.3, 0.8]}}'
-
-# $in filter
-bm tool search-notes "" --filter '{"priority": {"$in": ["high", "critical"]}}'
-```
-
-### `--tag` and `--status` — convenience shortcuts
-
-```bash
-bm tool search-notes "query" --tag security --tag oauth
-bm tool search-notes "" --status draft
-```
-
-### Combined example
-
-```bash
-bm tool search-notes "authentication" --tag security --meta status=draft --type spec
-```
-
-## Practical Examples
-
-### Example notes with custom frontmatter
-
-**`specs/auth-design.md`:**
-
-```markdown
----
-title: Auth Design
-type: spec
-tags: [security, oauth]
-status: in-progress
-priority: high
-confidence: 0.85
----
-
-# Auth Design
-
-## Observations
-- [decision] Use OAuth 2.1 with PKCE for all client types #security
-- [requirement] Token refresh must be transparent to the user
-
-## Relations
-- implements [[Security Requirements]]
-```
-
-**`specs/search-redesign.md`:**
-
-```markdown
----
-title: Search Redesign
-type: spec
-tags: [search, performance]
-status: draft
-priority: medium
-confidence: 0.6
----
-
-# Search Redesign
-
-## Observations
-- [goal] Sub-100ms search response times #performance
-- [approach] Hybrid FTS + vector retrieval
-
-## Relations
-- depends_on [[Database Schema]]
-```
-
-### Queries that find them
-
-```python
-# Find all in-progress specs
-await search_notes(metadata_filters={"status": "in-progress", "type": "spec"})
-# → Auth Design
-
-# Find high-confidence specs
-await search_notes(metadata_filters={"confidence": {"$gt": 0.7}})
-# → Auth Design (confidence: 0.85)
-
-# Find specs with priority high or medium
-await search_notes(metadata_filters={"priority": {"$in": ["high", "medium"]}})
-# → Auth Design, Search Redesign
-
-# Find specs in a confidence range
-await search_notes(metadata_filters={"confidence": {"$between": [0.5, 0.9]}})
-# → Auth Design (0.85), Search Redesign (0.6)
-
-# Find notes tagged with security
-await search_notes("tag:security")
-# → Auth Design
-
-# Combined: text search + metadata filter
-await search_notes("OAuth", metadata_filters={"status": "in-progress"})
-# → Auth Design
-```
-
-### CLI equivalents
-
-```bash
-bm tool search-notes "" --meta status=in-progress --type spec
-bm tool search-notes "" --filter '{"confidence": {"$gt": 0.7}}'
-bm tool search-notes "OAuth" --meta status=in-progress
-bm tool search-notes --tag security
-```

docs/post-v0.18.0-test-plan.md

@@ -1,344 +0,0 @@
-# Post-v0.18.0 Test Plan and Acceptance Criteria
-
-## Goal
-
-Define a complete validation plan for all major features merged after `v0.18.0`, combining:
-
-- Coverage-gap-driven automated tests
-- Real MCP server integration tests (no mocks for target flows)
-- Manual MCP verification via LLM-driven tool calls
-
-This plan is based on commits in `v0.18.0..HEAD` and the latest `just check` coverage output.
-
-## Scope Window
-
-- Start tag: `v0.18.0` (2026-01-28)
-- End: current `main`
-- Change volume: 12 feature commits + 14 bug-fix commits (+ release chores/hotfixes)
-
-## Execution Strategy
-
-1. Stabilize all feature-level acceptance criteria in automated tests first.
-2. Add black-box MCP integration tests for semantic search + schema (real server startup).
-3. Run manual MCP tool-call verification to confirm real UX and routing behavior.
-4. Re-run full gate: `just check` + targeted integration packs.
-
-## Global Quality Gates
-
-- Feature criteria below must all pass.
-- No regressions in existing suites.
-- Coverage improves in targeted low-coverage feature modules.
-- SQLite and Postgres parity for search/semantic features.
-
-## Priority Coverage Gaps (from latest run)
-
-These are the most important post-`v0.18.0` feature modules currently under-covered:
-
-- `src/basic_memory/mcp/tools/schema.py` (27%)
-- `src/basic_memory/mcp/clients/schema.py` (36%)
-- `src/basic_memory/mcp/tools/ui_sdk.py` (43%)
-- `src/basic_memory/mcp/tools/search.py` (73%)
-- `src/basic_memory/repository/postgres_search_repository.py` (63%)
-- `src/basic_memory/mcp/async_client.py` (82%)
-- `src/basic_memory/api/v2/routers/schema_router.py` (80%)
-
-## Feature Acceptance Criteria and Test Plan
-
-### 1) Schema System (`c97733d`) — DONE
-
-### Acceptance criteria
-
-- `schema_validate`, `schema_infer`, and `schema_diff` produce consistent outcomes across CLI/API/MCP for the same fixture set.
-- Strict validation fails deterministically on required-field/type violations.
-- Validation warnings are stable and machine-readable in non-strict mode.
-- Inference output is deterministic for unchanged input corpus.
-- Drift diff output is deterministic and identifies missing/extra/type-mismatch fields correctly.
-
-### Existing coverage anchor points
-
-- `tests/schema/*`
-- `tests/api/v2/test_schema_router.py`
-- `test-int/test_schema/*`
-
-### Gaps to close — DONE
-
-- ~~MCP schema tool branches (`src/basic_memory/mcp/tools/schema.py`)~~ — 18 tests in `tests/mcp/test_tool_schema.py`
-- ~~MCP schema client behavior (`src/basic_memory/mcp/clients/schema.py`)~~ — `tests/mcp/test_client_schema.py`
-- ~~Schema router error-path branches (`src/basic_memory/api/v2/routers/schema_router.py`)~~ — `tests/api/v2/test_schema_router.py`
-
-### Planned additions — DONE
-
-- ~~Add MCP tool tests for `schema_validate` strict + non-strict result shapes.~~ **DONE**
-- ~~Add MCP tool tests for `schema_infer` with explicit `entity_type` and inferred type fallback.~~ **DONE**
-- ~~Add MCP tool tests for `schema_diff` empty-diff and non-empty-diff paths.~~ **DONE**
-- ~~Add API tests for schema router invalid payload/edge error handling.~~ **DONE**
-- Add integration test that starts MCP server and calls schema tools end-to-end on fixture notes. — deferred to backlog item 4.
-
-### 2) Semantic Search (`0777879`, `1428d18`, `344e651`) — DONE
-
-### Acceptance criteria
-
-- `search_type=text|vector|hybrid` returns expected ranked results on canonical semantic corpus.
-- Missing semantic dependencies fail fast with actionable install guidance.
-- Reindex and provider/model changes produce valid vectors without dimension mismatch.
-- SQLite and Postgres produce equivalent behavior for semantic modes on the same dataset.
-- Generated-column migration path is valid on SQLite environments in use.
-
-### Existing coverage anchor points
-
-- `tests/repository/test_sqlite_vector_search_repository.py`
-- `tests/repository/test_postgres_search_repository.py`
-- `tests/services/test_semantic_search.py`
-- `tests/mcp/test_tool_search.py`
-- `test-int/test_search_performance_benchmark.py`
-
-### Gaps to close — DONE
-
-- ~~Uncovered Postgres vector/hybrid branches~~ — 20 tests in `tests/repository/test_postgres_search_repository_unit.py` + 5 integration tests in `test-int/semantic/test_semantic_coverage.py`
-- ~~MCP search semantic/output branches~~ — expanded `tests/mcp/test_tool_search.py`
-
-### Planned additions — DONE
-
-- ~~Expand Postgres repository tests for vector query composition edge cases.~~ **DONE**
-- ~~Expand Postgres repository tests for hybrid fusion ranking and pagination branches.~~ **DONE**
-- ~~Expand Postgres repository tests for embedding/provider error handling branches.~~ **DONE**
-- ~~Expand MCP search tool tests for vector/hybrid output formatting branches.~~ **DONE**
-- ~~Expand MCP search tool tests for semantic-disabled and missing-dependency failures.~~ **DONE**
-- Add MCP integration tests that start server and execute semantic `search_notes` tool calls. — deferred to backlog item 4.
-
-### Semantic search quality benchmarks (NEW)
-
-Full benchmark suite in `test-int/semantic/` covering 5 backend×provider combinations:
-- `sqlite-fts`, `sqlite-fastembed`, `postgres-fts`, `postgres-fastembed`, `postgres-openai`
-- Quality metrics: hit@1, recall@5, MRR@10 with per-query timing
-- Realistic corpus with cross-topic vocabulary overlap (240 notes, 4 topics)
-- Rich CLI viewer: `just semantic-report`
-- JSON artifact output: `just test-semantic-report`
-
-Key finding: **FastEmbed (384-d local ONNX) matches or exceeds OpenAI (1536-d) quality at 30x lower latency.** Recommending FastEmbed as default for both local and cloud deployments.
-
-### 3) Per-Project Local/Cloud Routing + API Key Auth (`d84708c`, `ed94877`, `312662f`) — DONE
-
-### Acceptance criteria
-
-- Project mode (`local`/`cloud`) persists and displays correctly.
-- Routing selects ASGI for local projects and HTTP+Bearer for cloud projects.
-- Cloud project without key fails with explicit remediation (`cloud set-key`/`cloud create-key`).
-- Resolution precedence is correct (factory > force-local > per-project cloud > global fallback > local).
-- Watch/sync only run for local projects.
-
-### Existing coverage anchor points
-
-- `tests/mcp/test_async_client_modes.py`
-- `tests/cli/test_project_set_cloud_local.py`
-- `tests/mcp/test_project_context.py`
-- `tests/test_project_resolver.py`
-- `tests/sync/test_watch_service_reload.py`
-
-### Gaps to close — DONE
-
-- ~~Cloud routing branch gaps in `src/basic_memory/mcp/async_client.py`~~ — expanded `tests/mcp/test_async_client_modes.py`
-
-### Planned additions — DONE
-
-- ~~Add branch-focused tests for all unresolved routing branches in `get_client()`.~~ **DONE**
-- Add MCP integration scenario with mixed local/cloud project config — deferred to backlog item 4.
-
-### 4) Project-Prefixed Permalinks + Memory URL Routing (`545804f`) — DONE
-
-### Acceptance criteria
-
-- Project-prefixed permalinks are generated consistently on create/update/import flows.
-- Memory URLs resolve to the correct project/entity even with duplicate note titles.
-- `read_note`, `search`, `build_context`, write/edit/move flows preserve project identity correctly.
-- Link resolution remains correct for context-aware wikilinks.
-
-### Existing coverage anchor points
-
-- `tests/utils/test_permalink_formatting.py`
-- `tests/mcp/test_tool_read_note.py`
-- `tests/mcp/test_tool_search.py`
-- `tests/services/test_context_service.py`
-- `test-int/mcp/test_read_note_integration.py`
-
-### Gaps to close
-
-- No major coverage alarm in report, but keep as regression-critical due broad impact surface.
-
-### Planned additions — DONE
-
-- ~~Add one integration test with colliding titles across two projects and assert URL routing invariants.~~ **DONE** — `test-int/mcp/test_permalink_collision_integration.py` (2 tests: collision across projects + memory:// URL routing with project prefix)
-
-### 5) MCP UI Variants + TUI Output (`8bc03d1`) — DONE
-
-### Acceptance criteria
-
-- UI resource variant selection (`tool-ui`, `vanilla`, `mcp-ui`) follows env configuration.
-- `search_notes` and `read_note` expose expected resource metadata for UI hosts.
-- `ascii`/`ansi` outputs are deterministic and stable for terminal clients.
-
-### Existing coverage anchor points
-
-- `tests/mcp/test_tool_contracts.py`
-- `test-int/mcp/test_output_format_json_integration.py`
-- `test-int/mcp/test_ui_sdk_integration.py`
-
-### Gaps to close — DONE
-
-- ~~`src/basic_memory/mcp/tools/ui_sdk.py` branch coverage~~ — `tests/mcp/test_ui_sdk.py`
-- ~~`src/basic_memory/mcp/ui/sdk.py` and `src/basic_memory/mcp/ui/templates.py` branch coverage~~ — `tests/mcp/test_ui_templates.py` + `tests/mcp/test_ui_resources.py`
-
-### Planned additions — DONE
-
-- ~~Add unit tests for UI SDK metadata generation and template selection branches.~~ **DONE** — 31 tests
-- ~~Add integration assertion for variant-specific resource URIs and metadata payload shape.~~ **DONE**
-
-### 6) Watch Command (`8df88e4`) — DONE
-
-### Acceptance criteria
-
-- `basic-memory watch` starts and processes create/update/delete events.
-- Watch restart/reload path does not duplicate watchers.
-- Cloud-mode projects are excluded from active watcher set.
-
-### Existing coverage anchor points
-
-- `tests/cli/test_watch.py`
-- `tests/sync/test_coordinator.py`
-- `tests/sync/test_watch_service_reload.py`
-
-### Planned additions — DONE
-
-- ~~Add one stress-style integration test for rapid file changes and watcher stability.~~ **DONE** — `tests/sync/test_watch_service_stress.py` (3 tests: 50-file batch, mixed add/modify/delete batch, rapid modifications to same file)
-
-### 7) CLI JSON Output (`a47c9c0`) — DONE
-
-### Acceptance criteria
-
-- `--format json` returns valid JSON with stable keys for success paths.
-- Error paths also return JSON-shaped output with correct non-zero exits.
-- Default human output remains unchanged.
-
-### Existing coverage anchor points
-
-- `tests/cli/test_cli_tool_json_output.py`
-- `test-int/cli/test_cli_tool_json_integration.py`
-
-### Planned additions — DONE
-
-- ~~Add one failure-path integration test per high-use tool command.~~ **DONE** — `test-int/cli/test_cli_tool_json_failure_integration.py` (4 tests: read-note not found, write-note missing content, write→read roundtrip, recent-activity empty project)
-
-### 8) Search/Edit and Metadata Fixes (`530cbac`, `f1d50c2`, `8838571`, `009e849`) — DONE
-
-### Acceptance criteria
-
-- Metadata filters produce consistent results on SQLite and Postgres.
-- `tag:` shorthand works alone and with mixed query terms.
-- Fast write/edit paths preserve `external_id` and metadata integrity.
-
-### Existing coverage anchor points
-
-- `tests/repository/test_metadata_filters.py`
-- `tests/repository/test_search_repository.py`
-- `tests/services/test_search_service.py`
-
-### Planned additions — DONE
-
-- ~~Add Postgres-specific metadata filter edge-case tests to mirror SQLite assertions exactly.~~ **DONE** — `tests/repository/test_metadata_filters_edge_cases.py` (6 tests: missing field, AND logic, contains single-element array, nested path missing intermediate, $gte/$lte boundaries, $between inclusive — all pass on both SQLite and Postgres)
-
-### 9) Compatibility and Hotfix Regression Pack (`c46d7a6`, `a0e754b`, `343a6e1`, `24ca5f6`, `e3ced49`, `8489a3d`, `b609c4e`, `f6e0a5b`, `7624a20`)
-
-### Acceptance criteria
-
-- Legacy endpoints required by older CLI versions function without `405` (`GET /projects/projects`, `POST /projects/projects`, `POST /projects/config/sync`).
-- Entity creation conflicts map to conflict status (not 500).
-- `recent_activity` prompt defaults are correct.
-- No spurious `metadata: {}` in serialized frontmatter.
-- Tigris/rclone uses global consistency headers for all transaction types.
-- `bm --version` fast path avoids heavy import path and remains responsive.
-- Default SQLite DB path is isolated by config dir.
-
-### Gaps to close
-
-- ~~Commits with no direct tests added (`c46d7a6`, `344e651`, `f6e0a5b`) need explicit regression tests.~~ **DONE**
-
-### Planned additions — DONE
-
-- ~~Add API compat test covering all legacy endpoint methods and payloads.~~ **DONE** — `test_legacy_v1_add_project_endpoint`, `test_legacy_v1_sync_config_endpoint`
-- ~~Add CLI fast-path test for `--version` import behavior/performance guard.~~ **DONE** — `test_bm_version_does_not_import_heavy_modules`
-- ~~Add empty metadata serialization regression test.~~ **DONE** — `test_schema_to_markdown_empty_metadata_no_metadata_key`
-- Add migration safety test for SQLite generated columns (`VIRTUAL` expectation) — deferred, low risk.
-
-## MCP Manual Verification Plan (LLM Tool Calls)
-
-Run after automated tests pass.
-
-### Setup
-
-- Start MCP server: `basic-memory mcp --transport stdio`
-- Use an MCP-capable client and issue tool calls directly.
-
-### Manual scenarios
-
-- Schema: call `schema_validate`, `schema_infer`, and `schema_diff` on known fixtures.
-- Schema: verify error and success payloads match acceptance criteria.
-- Semantic search: call `search_notes` with `search_type=text|vector|hybrid`.
-- Semantic search: verify ranking relevance on semantic fixture queries.
-- Routing: call tools with explicit project on mixed local/cloud setup.
-- Routing: verify success/failure paths with and without API key.
-- Permalink routing: read/write/search notes across projects with colliding titles.
-- Permalink routing: verify memory URL routing correctness.
-- UI/TUI: call `search_notes` and `read_note` with UI variants and `output_format=text|json`.
-- UI/TUI: verify payload/resource format and metadata completeness.
-
-## Implementation Backlog (Ordered)
-
-1. ~~Fill schema MCP/client/router coverage gaps.~~ **DONE** — 18 tests in `test_tool_schema.py` + `test_client_schema.py`
-2. ~~Fill semantic search MCP + Postgres repository gaps.~~ **DONE** — 20 tests in `test_postgres_search_repository_unit.py` + `test_tool_search.py`
-3. ~~Add compatibility regression tests (legacy endpoints, migration, version fast path).~~ **DONE** — 5 tests across 3 files (see below)
-4. ~~Add feature-level integration tests (permalinks, watch, CLI JSON, metadata filters).~~ **DONE** — 15 tests across 4 files (see items 4, 6, 7, 8 above)
-5. ~~Expand UI SDK and template branch tests.~~ **DONE** — 31 tests in `test_ui_templates.py` + `test_ui_sdk.py` + `test_ui_resources.py`
-6. ~~Run full gate and capture results in a short release readiness summary.~~ **DONE** — see results below
-
-### Full Gate Results (`just check`)
-
-| Phase | Result |
-|-------|--------|
-| lint | PASS |
-| format | PASS |
-| typecheck | PASS |
-| Unit tests (SQLite) | 1788 passed, 15 skipped |
-| Integration tests (SQLite) | 243 passed, 4 skipped, 10 deselected |
-| Unit tests (Postgres) | 1760 passed, 28 skipped |
-| Integration tests (Postgres) | 234 passed, 13 skipped, 10 deselected |
-
-**0 failures. 10 deselected = semantic benchmark tests (run separately via `just test-semantic`).**
-
-### Item 3 Details — Compatibility Regression Tests
-
-| Test | File | What it covers |
-|------|------|----------------|
-| `test_legacy_v1_add_project_endpoint` | `tests/api/v2/test_project_router.py` | POST `/projects/projects` legacy route reachable (idempotent path) |
-| `test_legacy_v1_sync_config_endpoint` | `tests/api/v2/test_project_router.py` | POST `/projects/config/sync` legacy route reachable |
-| `test_bm_version_does_not_import_heavy_modules` | `tests/cli/test_cli_exit.py` | `bm --version` fast path does not load `basic_memory.mcp` |
-| `test_schema_to_markdown_empty_metadata_no_metadata_key` | `tests/markdown/test_entity_parser_error_handling.py` | `schema_to_markdown()` with `entity_metadata={}` emits no `metadata:` key |
-| `test_legacy_v1_list_projects_endpoint` | `tests/api/v2/test_project_router.py` | (pre-existing) GET `/projects/projects` legacy route |
-
-**Suite totals after item 3: 1764 passed, 15 skipped, 0 failures.**
-
-## Suggested Commands
-
-- Full suite: `just check`
-- Fast loop: `just fast-check`
-- E2E consistency: `just doctor`
-- SQLite focused: `just test-sqlite`
-- Postgres focused: `just test-postgres`
-- Schema integration: `pytest test-int/test_schema -q`
-- Semantic + repo focus: `pytest tests/repository/test_postgres_search_repository.py tests/mcp/test_tool_search.py tests/services/test_semantic_search.py -q`
-- MCP integration focus: `pytest test-int/mcp -q`
-
-## Exit Criteria for This Plan
-
-- All feature acceptance criteria above are validated.
-- All identified high-priority coverage gaps are addressed or explicitly documented as intentional.
-- Manual MCP verification scenarios complete with no P0/P1 findings.

docs/releases/v0.19.0.md

@@ -1,318 +0,0 @@
-# v0.19.0 Release Notes
-
-## Overview
-
-v0.19.0 is a major release that introduces semantic vector search, a schema validation system,
-project-prefixed permalinks, per-project cloud routing, and a significant upgrade to FastMCP 3.0.
-It includes 90+ commits since v0.18.0 spanning new features, architectural improvements, and
-stability fixes across both SQLite and Postgres backends.
-
----
-
-## Major Features
-
-### Semantic Vector Search
-
-Full vector and hybrid search for SQLite (via sqlite-vec) and Postgres (via pgvector).
-
-- **Hybrid search mode** combines full-text search (FTS) with vector similarity for best results
-- **Score-based fusion** replaces RRF for hybrid ranking — `max(vec, fts) + 0.3 * min(vec, fts)` preserves dominant signals and rewards dual-source agreement (#577)
-- **Default search mode** is now `hybrid` when semantic search is enabled, `text` when disabled
-- Embedding providers: FastEmbed (local, default) or OpenAI API
-- Configurable similarity threshold via `semantic_min_similarity` (default 0.55)
-- Per-query `min_similarity` override on `search_notes` tool
-- Auto-backfill: existing entities get embeddings generated on first startup
-- Backend-specific distance-to-similarity conversion (cosine for SQLite, inner product for Postgres)
-- FTS fallback: if semantic dependencies are missing, search gracefully degrades to text-only
-- sqlite-vec knn `k` parameter capped at 4096 to prevent backend errors
-
-**Configuration:**
-```json
-{
-  "semantic_search_enabled": true,
-  "semantic_embedding_provider": "fastembed",
-  "semantic_embedding_model": "bge-small-en-v1.5",
-  "semantic_min_similarity": 0.55
-}
-```
-
-**Usage:**
-```
-search_notes("machine learning concepts", search_type="hybrid")
-search_notes("similar to my notes on coffee", search_type="vector")
-search_notes("exact phrase match", search_type="text")
-search_notes("broad search", min_similarity=0.3)  # lower threshold for more results
-```
-
-### Schema System
-
-Validate note structure against user-defined schemas with frontmatter-based rules.
-
-- Define schemas as YAML in note frontmatter with field types, required fields, and constraints
-- Frontmatter validation during sync — malformed notes get clear error messages
-- Schema inference from existing notes to bootstrap schemas from your content
-- Schema diff to compare two schemas and see changes
-- Available via MCP tools and CLI
-
-### Project-Prefixed Permalinks
-
-Permalinks now include the project name for unambiguous cross-project references.
-
-- Memory URLs like `memory://project-name/folder/note` route to the correct project
-- Existing non-prefixed permalinks continue to work (backwards compatible)
-- Controlled by `permalinks_include_project` config (default: true)
-- `build_context` and `search_notes` auto-detect project from URL prefix
-
-### Per-Project Cloud Routing
-
-Individual projects can be routed through the cloud while others stay local.
-
-- Set a project to cloud mode: `bm project set-cloud research`
-- Revert to local: `bm project set-local research`
-- Uses API key authentication: `bm cloud set-key bmc_abc123...`
-- MCP tools automatically route based on each project's mode
-- Local MCP server (`bm mcp`) still uses local routing for all projects by default
-- `--local` and `--cloud` CLI flags override per-command
-
-### Workspace Selection
-
-Cloud projects can target specific workspaces for multi-tenant environments.
-
-- `workspace` parameter on MCP tools for explicit workspace targeting
-- CLI workspace-aware project listing with `bm project list`
-- Spinner feedback while fetching cloud projects
-
----
-
-## New Tools and Capabilities
-
-### Dashboard (`bm project info`)
-
-`bm project info` now displays an htop-inspired compact dashboard with:
-
-- Horizontal bar charts for note types (top 5)
-- Embedding coverage bar with Unicode block characters
-- Colored status dots for at-a-glance health
-- `EmbeddingStatus` schema and `get_embedding_status()` service method for programmatic access
-
-### Unified Metadata Search
-
-`search_by_metadata` has been merged into `search_notes` — one tool for all searches.
-`query` is now optional, so you can search purely by frontmatter metadata.
-
-```
-search_notes(metadata_filters={"status": "in-progress"})
-search_notes(metadata_filters={"tags": ["security", "oauth"]})
-search_notes(metadata_filters={"priority": {"$in": ["high", "critical"]}})
-search_notes(metadata_filters={"schema.confidence": {"$gt": 0.7}})
-search_notes(tags=["security"])  # convenience shorthand
-search_notes(status="draft")     # convenience shorthand
-```
-
-### JSON Output Mode
-
-All MCP tools now support `output_format="json"` for machine-readable responses.
-
-- Default remains `"text"` for human-readable output (no breaking changes)
-- `build_context` defaults to `"json"` with slimmed payloads (redundant fields stripped)
-- CLI tool commands support `--format json` flag
-
-### `tag:` Search Shorthand
-
-Search by tag using convenient shorthand syntax.
-
-```
-search_notes("tag:security")
-search_notes("tag:coffee AND tag:brewing")
-```
-
-### Entity User Tracking
-
-Entities now track `created_by` and `last_updated_by` fields for attribution.
-
-### Improved Search Result Content (#609)
-
-Search results now surface more relevant context:
-
-- `matched_chunk_text` populated for FTS-only hybrid results (no more fallback to truncated content)
-- `TOP_CHUNKS_PER_RESULT` increased from 3 to 5, catching answers deeper in large notes (~2700 → ~4500 chars)
-- `CONTENT_DISPLAY_LIMIT` doubled from 2000 to 4000 chars for results without matched chunks
-
-### `write_note` Overwrite Guard (#632)
-
-`write_note` is now non-idempotent by default. If a note already exists, the tool returns an
-error instead of silently overwriting. Pass `overwrite=True` to replace, or use `edit_note`
-for incremental updates. Config option `write_note_overwrite_default` restores the old upsert
-behavior.
-
----
-
-## Architecture Changes
-
-### Score-Based Hybrid Fusion (#577)
-
-RRF (Reciprocal Rank Fusion) compressed all fused scores to ~0.016, destroying ranking
-differentiation. The new formula `max(vec, fts) + FUSION_BONUS * min(vec, fts)` preserves
-dominant signals and rewards dual-source agreement. Zero-score results now produce zero
-fused score instead of receiving a 0.1 weight floor.
-
-### FastMCP 3.0 Upgrade
-
-Upgraded from FastMCP 2.12.3 to 3.0.1.
-
-- Tool annotations (`readOnlyHint`, `openWorldHint`) for better client integration
-- Improved MCP protocol compliance
-- Better error handling and context management
-
-### Prompts Call MCP Tools Directly
-
-MCP prompts (`search`, `continue_conversation`) now call MCP tools directly instead of
-going through API endpoints. This fixes empty results in discovery mode and ensures prompts
-use the same resolution logic as tools (including LinkResolver fallback).
-
-### build_context LinkResolver Fallback
-
-`build_context` now falls back to LinkResolver when an exact permalink lookup returns empty.
-This uses the same 7-strategy resolution pipeline as `read_note`, so callers no longer get
-empty results for valid note identifiers that don't match exact permalinks.
-
-### Sync Handles Semantic Dependency Errors Gracefully
-
-When sqlite-vec or another embedding provider is unavailable, `sync_file` now catches
-`SemanticDependenciesMissingError` separately. The entity is created and FTS-indexed
-successfully — only vector embeddings are skipped, with a clear warning:
-
-```
-WARNING: Semantic search dependencies missing — vector embeddings skipped for path=note.md.
-Run 'bm reindex --embeddings' after resolving the dependency issue.
-```
-
-### Unified Project Path
-
-Cloud projects with bisync now store the local filesystem path in `path` (not the Docker
-container path). Config migration automatically promotes `local_sync_path` → `path` for
-existing configs.
-
----
-
-## CLI Improvements
-
-### Status and Doctor Default to Local Routing
-
-`bm status` and `bm doctor` now default to local routing since they scan the local filesystem.
-Previously, cloud-mode projects would route these commands to the cloud API, which returned
-Docker-internal paths that don't exist locally.
-
-### `--format json` for CLI Tool Commands
-
-All `bm tool` subcommands support `--format json` for machine-readable output, enabling
-integration with scripts and plugins.
-
-### `--json` for Top-Level CLI Commands
-
-Five additional CLI commands now support `--json` for machine-readable output:
-
-- `bm status --json` — sync report with new/modified/deleted/moved files and skipped files
-- `bm project list --json` — structured project list with name, paths, routing mode, and defaults
-- `bm schema validate --json` — validation report with per-note pass/fail, warnings, and errors
-- `bm schema infer --json` — field frequency analysis and suggested schema definition
-- `bm schema diff --json` — drift report with new fields, dropped fields, and cardinality changes
-
-This complements the existing `bm project info --json` and `bm tool --format json` support,
-making all major CLI commands scriptable for CI pipelines and automation.
-
-### Cloud Promo and Analytics
-
-- Cloud promo panel shown on first run or version bump with OSS discount code
-- Anonymous usage telemetry via Umami Cloud (promo/login funnel events only)
-- Opt out with `BASIC_MEMORY_NO_PROMOS=1`
-- No PII, no file contents, no per-command tracking
-- See [Telemetry](https://github.com/basicmachines-co/basic-memory#telemetry) in README
-
----
-
-## Bug Fixes
-
-- **#577**: RRF fusion compressed all hybrid scores to ~0.016, destroying ranking differentiation
-- **#582**: build_context returns empty results on valid note identifiers
-- **#575**: Remove hardcoded "main" default from default_project
-- **#595**: recent_activity dedup and pagination across MCP tools
-- **#593**: Backend-specific distance-to-similarity conversion
-- **#592**: Strip NUL bytes from content before PostgreSQL search indexing
-- **#562**: Use VIRTUAL instead of STORED columns in SQLite migration
-- **#558**: Add X-Tigris-Consistent headers to all rclone commands
-- **#541**: Handle EntityCreationError as conflict
-- **#536**: Stabilize metadata filters on Postgres
-- **#533**: Fix recent_activity prompt defaults
-- **#530**: Prevent spurious `metadata: {}` in frontmatter output
-- **#601**: Return matched chunk text in search results
-- **#606**: Accept `null` for `expected_replacements` in `edit_note`
-- **#579, #607**: Guard against closed streams in promo panel and missing vector tables on shutdown
-- **#609**: FTS-only hybrid results missing `matched_chunk_text`; content limits too conservative
-- **#631**: `build_context` related_results schema validation failure — replaced fragile `_slim_context()` stripping with Pydantic `exclude=True` field config
-- **#630**: Skip workspace resolution when client factory is active — prevents 401 errors in cloud MCP server mode
-- **#30**: `tag:` prefix query fails with hybrid search — moved tag prefix parsing to MCP tool level so it works with all search modes
-- **#31**: `search_notes` returns cluttered observation/relation-level results — now defaults to entity-level results
-- **#28**: `schema_infer` and `schema_diff` return raw Pydantic models as "undefined" in LLM output — added markdown formatters
-- Fix `schema_validate` identifier resolution (now uses LinkResolver) and text rendering (markdown formatter)
-- **#634**: `schema_validate` and `schema_diff` use stale database metadata instead of reading schema definitions from file — now reads frontmatter directly from the file with fallback to database metadata
-- Fix `Post(**metadata)` crash when frontmatter contains `content` or `handler` keys
-- Fix list-valued frontmatter fields (`title`, `type`) crashing on `.strip()` — now coerced to strings
-- Cap sqlite-vec knn `k` parameter at 4096 to prevent backend errors
-- Parameterize SQL queries in search repository type filters
-- Double-default display in project list
-- `ensure_frontmatter_on_sync` default changed to `True`
-- Status/doctor commands fail with cloud-mode projects (Docker path error)
-- Prompts return "0 projects" in discovery mode
-
----
-
-## Security
-
-- Upgrade `cryptography` for CVE advisory
-- Upgrade `python-multipart` for security advisory
-
----
-
-## Internal / Developer
-
-- **#598**: Upgrade FastMCP 2.12.3 → 3.0.1 with tool annotations
-- **#594**: Add `ty` as supplemental type checker
-- **#538**: Add fast feedback loop tooling (`just fast-check`, `just doctor`, `just testmon`)
-- **#600**: Rename `entity_type` to `note_type` for consistency
-- **#596**: Fix CLI runtime defects and audit regressions
-- CLI refactoring and workspace-aware cloud project listing
-- Split and speed up PR test matrix in CI
-- Fix CI: collect coverage from test jobs instead of re-running all tests
-- Create `search_vector_chunks` in test fixtures for Postgres compatibility
-
----
-
-## Configuration Changes
-
-| Setting | Old Default | New Default | Notes |
-|---------|-------------|-------------|-------|
-| `semantic_search_enabled` | `false` | `true` | Semantic search on by default |
-| `ensure_frontmatter_on_sync` | `false` | `true` | Frontmatter added during sync |
-| `permalinks_include_project` | `false` | `true` | Project prefix in permalinks |
-
----
-
-## Upgrade Notes
-
-- **Semantic search dependencies** are now included by default. If sqlite-vec fails to load,
-  search gracefully falls back to FTS. Run `bm reindex --embeddings` to generate embeddings
-  for existing content.
-- **Hybrid search scoring** has changed from RRF to score-based fusion. Search result ordering
-  may differ — results should be more accurate with better score differentiation.
-- **`search_by_metadata`** is removed as a standalone tool. Use `search_notes` with
-  `metadata_filters` instead (same parameters, same behavior).
-- **Project-prefixed permalinks** are enabled by default. Existing notes keep their current
-  permalinks until modified. Set `permalinks_include_project: false` to disable.
-- **Frontmatter on sync** is now enabled by default. Files without frontmatter will have it
-  added on next sync. Set `ensure_frontmatter_on_sync: false` to preserve old behavior.
-- **Config migration** runs automatically for cloud projects with bisync — `local_sync_path`
-  is promoted to `path` so filesystem operations work correctly.
-- **`write_note` is no longer idempotent** — calls to `write_note` for existing notes now
-  return an error unless `overwrite=True` is passed. Use `edit_note` for incremental changes,
-  or set `write_note_overwrite_default: true` in config to restore the old behavior.

docs/semantic-search-test-log.md

@@ -1,209 +0,0 @@
-# Semantic Search Manual Test Log
-
-## Overview
-
-Manual test session for semantic (vector) search on the main project.
-- Date: 2026-02-15
-- Database: ~/.basic-memory/memory.db (SQLite)
-- Entities: 456 embedded, 2714 vector chunks
-- Search index: 2390 FTS entries
-- Embedding model: default (384-dim, sqlite-vec)
-
-## Test Plan
-
-1. **Search Type Routing** — verify vector/hybrid/text dispatch, invalid search_type handling
-2. **Conceptual Queries** — natural language where vector should beat FTS
-3. **Keyword Queries** — exact terms where FTS should be strong
-4. **Hybrid Ranking** — queries where both FTS and vector contribute
-5. **Result Types** — entities, observations, relations in vector results
-6. **Filters + Vector** — combine vector with types/entity_types/after_date
-7. **Edge Cases** — short queries, long queries, empty, special chars, no-match
-8. **Pagination** — page > 1, page_size respected
-
----
-
-## Test Results
-
-### Test 1: Search Type Routing
-
-#### 1a: search_type="semantic" (invalid value)
-- **Input:** query="how does the knowledge graph work", search_type="semantic"
-- **Expected:** error or explicit fallback
-- **Actual:** Silently falls through to text search (else branch in search.py:430)
-- **Verdict:** BUG — should either be a recognized alias for "vector" or return an error
-
-#### 1b: search_type="vector"
-- **Input:** query="keeping AI context between sessions", search_type="vector"
-- **Actual:** 5 results, scores ~0.58-0.59, found "Maintaining context across conversation boundaries" observation
-- **Verdict:** PASS
-
-#### 1c: search_type="text" with conceptual query
-- **Input:** query="keeping AI context between sessions", search_type="text"
-- **Actual:** 0 results (no exact keyword match)
-- **Verdict:** PASS (expected — FTS requires token overlap)
-
-#### 1d: search_type="hybrid" with conceptual query
-- **Input:** query="keeping AI context between sessions", search_type="hybrid"
-- **Actual:** 5 results, same ranking as vector (FTS contributed nothing here)
-- **Verdict:** PASS
-
-#### 1e: search_type="text" with keyword query
-- **Input:** query="OAuth authentication", search_type="text"
-- **Actual:** 3 results — AUTH.md Supabase OAuth, OAuth Rip-and-Replace, OAuth Integration Analysis
-- **Verdict:** PASS
-
-#### 1f: search_type="vector" with keyword query
-- **Input:** query="OAuth authentication", search_type="vector"
-- **Actual:** Same top results as text (keyword-rich content also scores well in vector space)
-- **Verdict:** PASS
-
----
-
-### Test 2: Conceptual Queries (vector advantage)
-
-#### 2a: Natural language question
-- **Input:** query="why do AI assistants forget things", search_type="vector"
-- **Actual:** 5 results — Manual Testing Session, "Balance security and usability" observation, "Tools should match thought patterns" observation. Scores ~0.56-0.57
-- **Vector advantage:** Found conceptually related content despite no exact keyword overlap
-- **Verdict:** PASS
-
-#### 2b: Same query, text search
-- **Input:** query="why do AI assistants forget things", search_type="text"
-- **Actual:** 1 result — "What is Basic Memory?" (likely matched on "AI" token)
-- **Verdict:** PASS (demonstrates vector advantage — text barely matched)
-
-#### 2c: Domain concept with no jargon
-- **Input:** query="pricing strategy for cloud product", search_type="vector"
-- **Actual:** 3 results — SPEC-16 MCP Cloud Service Consolidation, knowledge architecture observation, Visual Knowledge Spaces relation. Scores ~0.56-0.57
-- **Verdict:** PASS (found cloud-related content conceptually)
-
-#### 2d: Technical concept, long query
-- **Input:** query="SQLite performance optimization WAL mode concurrent writes", search_type="vector"
-- **Actual:** 3 results — SPEC-11 API Performance Optimization, Real-Time Updates with WebSockets, marketing status update. Scores ~0.55-0.58
-- **Verdict:** PASS (found performance-related content)
-
----
-
-### Test 3: Keyword Queries (FTS strength)
-
-#### 3a: Exact term match — "OAuth authentication"
-- **Text:** 3 results with high relevance (exact matches in titles)
-- **Vector:** Same top results (keyword overlap helps vector too)
-- **Verdict:** PASS — FTS and vector converge on keyword-rich queries
-
-#### 3b: "OAuth" single keyword, hybrid mode
-- **Input:** query="OAuth", search_type="hybrid"
-- **Actual:** 5 results — Basic Memory Coding Guide, AI Collaboration Examples, SPEC-18, daily note, Manual Testing Session. FTS + vector blended. Scores ~0.016-0.032
-- **Note:** Top hybrid result is "Basic Memory Coding Guide" not an OAuth-specific doc — suggests hybrid scoring may dilute strong FTS matches
-- **Verdict:** PASS but hybrid ranking questionable for single-keyword queries
-
----
-
-### Test 4: Hybrid Ranking
-
-#### 4a: Hybrid vs vector on "OAuth authentication"
-- **Hybrid with entity_types=["entity"]:** 5 results — RLS Implementation Lessons, Cloud Readiness Assessment, AUTH.md OAuth, Core Service Implementation, OAuth Rip-and-Replace. Scores ~0.016-0.023
-- **Vector with entity_types=["entity"]:** 5 results — Core Service Implementation, SPEC-13 CLI Auth, Coding Guide, Authentication Service, ADR Production Auth. Scores ~0.55-0.60
-- **Observation:** Hybrid surfaces different top results than vector-only. Hybrid found RLS and Cloud Readiness docs that vector didn't prioritize. Different ranking is expected from RRF fusion.
-- **Verdict:** PASS — hybrid produces meaningfully different ranking
-
----
-
-### Test 5: Result Types
-
-#### 5a: Vector returns all result types
-- **Input:** query="keeping AI context between sessions", search_type="vector"
-- **Entities:** SPEC-18 AI Memory Management Tool (type=entity)
-- **Relations:** Prompt Builder integrates_with (type=relation)
-- **Observations:** "Translation layer is key" (type=observation), "Maintaining context across conversation boundaries" (type=observation)
-- **Verdict:** PASS — all three types appear in vector results
-
-#### 5b: Observations carry metadata
-- **Observation result:** category="challenge", content="Maintaining context across conversation boundaries", from_entity="research/ai-knowledge-management-research"
-- **Verdict:** PASS — category, content, from_entity, tags all present
-
-#### 5c: Relations carry link info
-- **Relation result:** relation_type="integrates_with", from_entity="development/features/prompt-builder...", to_entity (present but truncated in some)
-- **Verdict:** PASS — relation metadata present
-
----
-
-### Test 6: Filters + Vector Search
-
-#### 6a: entity_types=["entity"] with vector
-- **Input:** query="OAuth authentication", search_type="vector", entity_types=["entity"]
-- **Actual:** 5 results, all type="entity" (Core Service Implementation, SPEC-13, Coding Guide, Authentication Service, ADR Auth)
-- **Verdict:** PASS — filter correctly restricts to entities only
-
-#### 6b: types=["note"] with vector
-- **Input:** query="OAuth authentication", search_type="vector", types=["note"]
-- **Actual:** Same 5 results (all have entity_type="note" in metadata)
-- **Verdict:** PASS — types filter works with vector search
-
-#### 6c: after_date with vector
-- **Input:** query="OAuth authentication", search_type="vector", after_date="2025-06-01"
-- **Actual:** 3 results — Core Service Implementation, Cloud Web App analysis observation, SPEC-13. Filtered out older OAuth docs.
-- **Verdict:** PASS — date filter applied correctly
-
-#### 6d: entity_types=["entity"] with hybrid
-- **Input:** query="OAuth authentication", search_type="hybrid", entity_types=["entity"]
-- **Actual:** 5 results, all type="entity" — RLS lessons, Cloud Readiness, AUTH.md OAuth, Core Service, OAuth Rip-and-Replace
-- **Verdict:** PASS — filter works with hybrid mode too
-
-#### 6e: types=["entity"] with vector (WRONG filter name)
-- **Input:** query="OAuth authentication", search_type="vector", types=["entity"]
-- **Actual:** 0 results
-- **Note:** `types` filters by entity_type metadata (e.g., "note", "person"), NOT by SearchItemType. Using types=["entity"] looks for entity_type="entity" which few/no notes have. This is a UX confusion point — the param names are ambiguous.
-- **Verdict:** PASS (correct behavior) but USABILITY ISSUE — easy to confuse types vs entity_types
-
----
-
-### Test 7: Edge Cases
-
-#### 7a: Single character query
-- **Input:** query="x", search_type="vector"
-- **Actual:** 3 results — "Self-contained application bundle" observation, Non-Markdown File Support relation, quick-win-tools entity. Scores ~0.57-0.59
-- **Note:** Single character still produces an embedding and returns results. Quality is low/random as expected.
-- **Verdict:** PASS (no crash, returns results)
-
-#### 7b: Whitespace-only query
-- **Input:** query="   ", search_type="vector"
-- **Actual:** 0 results
-- **Verdict:** PASS (handled gracefully — _check_vector_eligible strips and rejects empty)
-
-#### 7c: Query with no relevant content
-- **Input:** query="quantum computing blockchain", search_type="vector"
-- **Actual:** 3 results — Inter-Agent Communication relation, Self-contained bundle observation, JSON-LD interop observation. Scores ~0.54
-- **Note:** Still returns results because vector search always finds nearest neighbors. Scores are lower (~0.54) than relevant queries (~0.58-0.60). No relevance threshold applied.
-- **Verdict:** PASS (expected behavior) but NOTE — no relevance cutoff means irrelevant queries always return something
-
----
-
-### Test 8: Pagination
-
-#### 8a: Vector search page 2
-- **Input:** query="keeping AI context between sessions", search_type="vector", page=2, page_size=3
-- **Actual:** 3 results on page 2, current_page=2. Different results from page 1. Top: "Maintaining context across conversation boundaries" observation (score 0.587)
-- **Note:** Interestingly, page 2 had a higher-scoring result than some page 1 results. This may indicate pagination doesn't sort globally — it might be paginating within a pre-scored set.
-- **Verdict:** PASS (pagination works) but POSSIBLE ISSUE — result ordering across pages needs investigation
-
----
-
-## Summary
-
-### Passing Tests: 20/21
-
-### Bugs Found
-1. **search_type="semantic" silently falls through** (Test 1a) — Invalid search_type values fall to the `else` branch and default to text search without any warning. Should either alias "semantic" to "vector" or raise an error.
-
-### Usability Issues
-2. **types vs entity_types confusion** (Test 6e) — `types` filters by entity_type metadata (note, person, etc.) while `entity_types` filters by SearchItemType (entity, observation, relation). The naming is ambiguous and easy to mix up.
-3. **No relevance threshold** (Test 7c) — Vector search always returns nearest neighbors even for completely irrelevant queries. Consider adding a minimum score threshold or at least documenting expected score ranges.
-4. **Hybrid ranking for single keywords** (Test 3b) — Hybrid mode on simple keyword queries produced less intuitive rankings than pure FTS or pure vector. The RRF fusion may dilute strong FTS signals.
-
-### Observations
-- Vector search successfully finds conceptually related content that FTS misses entirely
-- Score ranges: relevant queries ~0.56-0.60, irrelevant queries ~0.54 (narrow spread)
-- All three result types (entity, observation, relation) appear correctly in vector results
-- Filters (entity_types, types, after_date) all work correctly with vector and hybrid modes
-- Pagination works but cross-page ordering may need investigation

docs/semantic-search.md

@@ -1,270 +0,0 @@
-# Semantic Search
-
-This guide covers Basic Memory's semantic (vector) search feature, which adds meaning-based retrieval alongside the existing full-text search.
-
-## Overview
-
-Basic Memory's search supports both full-text search (FTS) and semantic retrieval. Semantic search adds vector embeddings that capture the *meaning* of your content, enabling:
-
-- **Paraphrase matching**: Find "authentication flow" when searching for "login process"
-- **Conceptual queries**: Search for "ways to improve performance" and find notes about caching, indexing, and optimization
-- **Hybrid retrieval**: Combine the precision of keyword search with the recall of semantic similarity
-
-Semantic search is enabled by default when semantic dependencies are available at runtime. It works on both SQLite (local) and Postgres (cloud) backends.
-
-## Installation
-
-Semantic search dependencies (fastembed, sqlite-vec, openai) are included in the default `basic-memory` install.
-
-```bash
-pip install basic-memory
-```
-
-You can always override with `BASIC_MEMORY_SEMANTIC_SEARCH_ENABLED=true|false`.
-
-### Platform Compatibility
-
-| Platform | FastEmbed (local) | OpenAI (API) |
-|---|---|---|
-| macOS ARM64 (Apple Silicon) | Yes | Yes |
-| macOS x86_64 (Intel Mac) | No — see workaround below | Yes |
-| Linux x86_64 | Yes | Yes |
-| Linux ARM64 | Yes | Yes |
-| Windows x86_64 | Yes | Yes |
-
-#### Intel Mac Workaround
-
-The default install includes FastEmbed, which depends on ONNX Runtime. ONNX Runtime dropped Intel Mac (x86_64) wheels starting in v1.24, so install with a compatible ONNX Runtime pin first:
-
-```bash
-pip install basic-memory 'onnxruntime<1.24'
-```
-
-After installation, Intel Mac users have two runtime options:
-
-**Option 1: Use OpenAI embeddings (recommended)**
-
-```bash
-export BASIC_MEMORY_SEMANTIC_SEARCH_ENABLED=true
-export BASIC_MEMORY_SEMANTIC_EMBEDDING_PROVIDER=openai
-export OPENAI_API_KEY=sk-...
-```
-
-**Option 2: Use FastEmbed locally**
-
-Keep the same pinned installation and use FastEmbed (default provider):
-
-```bash
-export BASIC_MEMORY_SEMANTIC_SEARCH_ENABLED=true
-export BASIC_MEMORY_SEMANTIC_EMBEDDING_PROVIDER=fastembed
-```
-
-## Quick Start
-
-1. Install Basic Memory:
-
-```bash
-pip install basic-memory
-```
-
-2. (Optional) Explicitly enable semantic search:
-
-```bash
-export BASIC_MEMORY_SEMANTIC_SEARCH_ENABLED=true
-```
-
-3. Build vector embeddings for your existing content:
-
-```bash
-bm reindex --embeddings
-```
-
-4. Search using semantic modes:
-
-```python
-# Pure vector similarity
-search_notes("login process", search_type="vector")
-
-# Hybrid: combines FTS precision with vector recall (recommended)
-search_notes("login process", search_type="hybrid")
-
-# Explicit full-text search
-search_notes("login process", search_type="text")
-```
-
-## Configuration Reference
-
-All settings are fields on `BasicMemoryConfig` and can be set via environment variables (prefixed with `BASIC_MEMORY_`).
-
-| Config Field | Env Var | Default | Description |
-|---|---|---|---|
-| `semantic_search_enabled` | `BASIC_MEMORY_SEMANTIC_SEARCH_ENABLED` | Auto (`true` when semantic deps are available) | Enable semantic search. Required before vector/hybrid modes work. |
-| `semantic_embedding_provider` | `BASIC_MEMORY_SEMANTIC_EMBEDDING_PROVIDER` | `"fastembed"` | Embedding provider: `"fastembed"` (local) or `"openai"` (API). |
-| `semantic_embedding_model` | `BASIC_MEMORY_SEMANTIC_EMBEDDING_MODEL` | `"bge-small-en-v1.5"` | Model identifier. Auto-adjusted per provider if left at default. |
-| `semantic_embedding_dimensions` | `BASIC_MEMORY_SEMANTIC_EMBEDDING_DIMENSIONS` | Auto-detected | Vector dimensions. 384 for FastEmbed, 1536 for OpenAI. Override only if using a non-default model. |
-| `semantic_embedding_batch_size` | `BASIC_MEMORY_SEMANTIC_EMBEDDING_BATCH_SIZE` | `64` | Number of texts to embed per batch. |
-| `semantic_vector_k` | `BASIC_MEMORY_SEMANTIC_VECTOR_K` | `100` | Candidate count for vector nearest-neighbour retrieval. Higher values improve recall at the cost of latency. |
-
-## Embedding Providers
-
-### FastEmbed (default)
-
-FastEmbed runs entirely locally using ONNX models — no API key, no network calls, no cost.
-
-- **Model**: `BAAI/bge-small-en-v1.5`
-- **Dimensions**: 384
-- **Tradeoff**: Smaller model, fast inference, good quality for most use cases
-
-```bash
-# Install basic-memory and enable semantic search
-pip install basic-memory
-export BASIC_MEMORY_SEMANTIC_SEARCH_ENABLED=true
-```
-
-### OpenAI
-
-Uses OpenAI's embeddings API for higher-dimensional vectors. Requires an API key.
-
-- **Model**: `text-embedding-3-small`
-- **Dimensions**: 1536
-- **Tradeoff**: Higher quality embeddings, requires API calls and an OpenAI key
-
-```bash
-export BASIC_MEMORY_SEMANTIC_SEARCH_ENABLED=true
-export BASIC_MEMORY_SEMANTIC_EMBEDDING_PROVIDER=openai
-export OPENAI_API_KEY=sk-...
-```
-
-When switching from FastEmbed to OpenAI (or vice versa), you must rebuild embeddings since the vector dimensions differ:
-
-```bash
-bm reindex --embeddings
-```
-
-## Search Modes
-
-### `text` (default)
-
-Full-text keyword search using FTS5 (SQLite) or tsvector (Postgres). Supports boolean operators (`AND`, `OR`, `NOT`), phrase matching, and prefix wildcards.
-
-```python
-search_notes("project AND planning", search_type="text")
-```
-
-This is the existing default and does not require semantic search to be enabled.
-
-### `vector`
-
-Pure semantic similarity search. Embeds your query and finds the nearest content vectors. Good for conceptual or paraphrase queries where exact keywords may not appear in the content.
-
-```python
-search_notes("how to speed up the app", search_type="vector")
-```
-
-Returns results ranked by cosine similarity. Individual observations and relations surface as first-class results, not collapsed into parent entities.
-
-### `hybrid`
-
-Combines FTS and vector results using score-based fusion. This is generally the best mode when you want both keyword precision and semantic recall.
-
-```python
-search_notes("authentication security", search_type="hybrid")
-```
-
-Score-based fusion uses the formula `max(vec, fts) + bonus * min(vec, fts)` to preserve the dominant signal while rewarding results found by both methods.
-
-### When to Use Which
-
-| Mode | Best For |
-|---|---|
-| `text` | Exact keyword matching, boolean queries, tag/category searches |
-| `vector` | Conceptual queries, paraphrase matching, exploratory searches |
-| `hybrid` | General-purpose search combining precision and recall |
-
-## The Reindex Command
-
-The `bm reindex` command rebuilds search indexes without dropping the database.
-
-```bash
-# Rebuild everything (FTS + embeddings if semantic is enabled)
-bm reindex
-
-# Only rebuild vector embeddings
-bm reindex --embeddings
-
-# Only rebuild the full-text search index
-bm reindex --search
-
-# Target a specific project
-bm reindex -p my-project
-```
-
-### When You Need to Reindex
-
-- **Upgrade note**: Migration now performs a one-time automatic embedding backfill on upgrade.
-- **Manual enable case**: If you explicitly had `semantic_search_enabled=false` and then turn it on
-- **Provider change**: After switching between `fastembed` and `openai`
-- **Model change**: After changing `semantic_embedding_model`
-- **Dimension change**: After changing `semantic_embedding_dimensions`
-
-The reindex command shows progress with embedded/skipped/error counts:
-
-```
-Project: main
-  Building vector embeddings...
-  ✓ Embeddings complete: 142 entities embedded, 0 skipped, 0 errors
-
-Reindex complete!
-```
-
-## How It Works
-
-### Chunking
-
-Each entity in the search index is split into semantic chunks before embedding:
-
-- **Headers**: Markdown headers (`#`, `##`, etc.) start new chunks
-- **Bullets**: Each bullet item (`-`, `*`) becomes its own chunk for granular fact retrieval
-- **Prose sections**: Non-bullet text is merged up to ~900 characters per chunk
-- **Long sections**: Oversized content is split with ~120 character overlap to preserve context at boundaries
-
-Each search index item type (entity, observation, relation) is chunked independently, so observations and relations are embeddable as discrete facts.
-
-### Deduplication
-
-Each chunk has a `source_hash` (SHA-256 of the chunk text). On re-sync, unchanged chunks skip re-embedding entirely. This makes incremental updates fast — only modified content triggers API calls or model inference.
-
-### Hybrid Fusion
-
-Hybrid search uses score-based fusion to merge FTS and vector results:
-
-1. Run FTS search to get keyword-ranked results; normalize scores to [0, 1]
-2. Run vector search to get similarity-ranked results (already [0, 1])
-3. For each result, compute: `fused = max(vec_score, fts_score) + 0.3 * min(vec_score, fts_score)`
-4. Sort by fused score
-
-The dominant signal (whichever source scored higher) is preserved, and dual-source agreement adds a bonus. Unlike rank-based fusion, this approach retains score magnitude — a strong vector match stays strong even without an FTS hit.
-
-### Observation-Level Results
-
-Vector and hybrid modes return individual observations and relations as first-class search results, not just parent entities. This means a search for "water temperature for brewing" can surface the specific observation about 205°F without returning the entire "Coffee Brewing Methods" entity.
-
-## Database Backends
-
-### SQLite (local)
-
-- **Vector storage**: [sqlite-vec](https://github.com/asg017/sqlite-vec) virtual table
-- **Table creation**: At runtime when semantic search is first used — no migration needed
-- **Embedding table**: `search_vector_embeddings` using `vec0(embedding float[N])` where N is the configured dimensions
-- **Chunk metadata**: `search_vector_chunks` table stores chunk text, keys, and source hashes
-
-The sqlite-vec extension is loaded per-connection. Vector tables are created lazily on first use.
-
-### Postgres (cloud)
-
-- **Vector storage**: [pgvector](https://github.com/pgvector/pgvector) with HNSW indexing
-- **Chunk metadata table**: Created via Alembic migration (`search_vector_chunks` with `BIGSERIAL` primary key)
-- **Embedding table**: `search_vector_embeddings` created at runtime (dimension-dependent, same pattern as SQLite)
-- **Index**: HNSW index on the embedding column for fast approximate nearest-neighbour queries
-
-The Alembic migration creates the dimension-independent chunks table. The embeddings table and HNSW index are deferred to runtime because they depend on the configured vector dimensions.

docs/specs/SPEC-LOCAL-PLUS-PUBLISH.md

@@ -1,225 +0,0 @@
-# SPEC-LOCAL-PLUS-PUBLISH: Local+ Published Notes and Privacy Tiers
-
-**Status:** Draft  
-**Date:** 2026-02-14  
-**Owner:** Basic Memory
-
-## Summary
-
-Add a paid Local+ feature that lets users publish selected notes to shareable URLs while keeping the
-main knowledge base local-first. Use this as a product wedge for users who do not want full cloud
-hosting but do want collaboration and distribution features.
-
-This spec also captures a practical position on "zero knowledge" for Local+.
-
-## Context
-
-Basic Memory already has strong local-first primitives and optional cloud routing/sync. A recurring
-request is:
-
-- keep knowledge local by default,
-- pay for selective value-add,
-- share specific outputs externally.
-
-Published Notes fits this model: explicit per-note opt-in, reversible, and easy to understand.
-
-## Goals
-
-1. Provide an Obsidian Publish-style sharing experience for selected notes.
-2. Keep local markdown files as source of truth.
-3. Make sharing compatible with current cloud/auth/billing primitives.
-4. Define clear Local+ packaging that does not degrade OSS local workflows.
-5. Document zero-knowledge constraints so product decisions are explicit.
-
-## Non-Goals
-
-1. Full hosted editing for all notes (Cloud Full remains separate).
-2. Public website builder/CMS features.
-3. Strict cryptographic zero-knowledge server processing for MCP/search in v1.
-
-## Local+ Feature Catalog (Sellable)
-
-Core Local+ candidates:
-
-1. Published Notes (share URL, revoke, expiry, password).
-2. Snapshot Time Machine (point-in-time restore for local projects).
-3. Recovery Drill Reports (automated restore verification).
-4. Device/API Key Governance (per-device keys, revocation, audit trail).
-5. BYO Storage Orchestration (managed setup for user-owned object storage).
-6. Semantic Boost Add-on (higher quality retrieval options while files remain source-of-truth).
-
-Team-oriented add-ons:
-
-1. Team-owned shared links and domain branding.
-2. Role-based publish permissions.
-3. Shared workspace policies for what can be published.
-
-## Proposed MVP: Published Notes
-
-### User Experience
-
-Per note actions:
-
-1. Publish.
-2. Unpublish.
-3. Copy URL.
-4. Regenerate URL.
-5. Set visibility and controls.
-
-Controls:
-
-1. Visibility: `unlisted` (default) or `public`.
-2. Optional password gate.
-3. Optional expiration datetime.
-4. Optional "disable indexing" flag for public mode.
-
-Behavior:
-
-1. Source note remains local markdown.
-2. Publish is explicit opt-in per note.
-3. Unpublish removes public access immediately.
-4. Republish creates a new URL token unless user chooses to keep current URL.
-
-### URL Model
-
-1. Unlisted share URL: high-entropy token path.
-2. Public URL: slug path (optional, later phase).
-3. Team plans can support custom domain mapping in later phase.
-
-### Content Model
-
-v1 published page includes:
-
-1. Rendered markdown body.
-2. Optional metadata (title, updated_at).
-
-v1 excludes:
-
-1. Full graph traversal expansion.
-2. Related note auto-discovery on public pages.
-
-### Sync Model
-
-1. Local file remains canonical.
-2. Publish stores a rendered snapshot plus metadata in cloud.
-3. Update path:
-   - manual "update published version", or
-   - optional auto-update on note change (plan-gated).
-
-## Architecture (v1)
-
-### High-Level Flow
-
-1. Client selects a note to publish.
-2. Client sends publish request with note identifier and policy.
-3. Service resolves note content (local sync artifact or explicit upload payload).
-4. Service stores published artifact and returns share URL.
-
-### Data Model
-
-`published_notes`
-
-1. `id` (uuid)
-2. `tenant_id` or `workspace_id`
-3. `project_id`
-4. `entity_permalink` (or stable external_id)
-5. `share_token` (hashed in DB)
-6. `visibility` (`unlisted`|`public`)
-7. `password_hash` (nullable)
-8. `expires_at` (nullable)
-9. `is_active`
-10. `published_content` (rendered snapshot or reference)
-11. `published_at`
-12. `updated_at`
-
-### API Shape (Draft)
-
-1. `POST /api/published-notes`
-2. `GET /api/published-notes`
-3. `GET /api/published-notes/{id}`
-4. `PATCH /api/published-notes/{id}`
-5. `DELETE /api/published-notes/{id}` (unpublish)
-6. `POST /api/published-notes/{id}/regenerate-url`
-7. `GET /p/{token}` (public resolver)
-
-### CLI Shape (Draft)
-
-1. `bm cloud publish <identifier>`
-2. `bm cloud publish list`
-3. `bm cloud publish update <id>`
-4. `bm cloud publish unpublish <id>`
-5. `bm cloud publish rotate-url <id>`
-
-### Security
-
-1. Default to unlisted URLs.
-2. Store only hashed share tokens.
-3. Passwords hashed server-side.
-4. Enforce expiration at request time.
-5. Log publish/unpublish/rotate events for auditability.
-
-## Packaging and Pricing Direction
-
-Suggested split:
-
-1. OSS Local: no publish URLs.
-2. Local+ Solo: publish URLs + snapshots + recovery.
-3. Local+ Team: solo features + team governance and branding.
-4. Cloud Full: hosted app + full cloud workflows.
-
-Key message:
-"Keep everything local. Publish only what you choose."
-
-## Rollout Plan
-
-1. Phase 1: Unlisted publish URLs + unpublish + regenerate URL.
-2. Phase 2: Password/expiry controls.
-3. Phase 3: Auto-update on note change and basic analytics.
-4. Phase 4: Team branding/domains/policies.
-
-## Zero-Knowledge Position
-
-### Strict Zero-Knowledge Definition
-
-Strict zero-knowledge means the server cannot decrypt note content at all.
-
-### Why This Conflicts with MCP and Search
-
-If server cannot decrypt:
-
-1. MCP tool execution against cloud content cannot read/write semantic content.
-2. Full-text search cannot index plaintext content.
-3. Semantic/vector search cannot generate or query embeddings on plaintext.
-4. Server-side relation resolution and context building become severely limited.
-
-This matches earlier findings: strict zero-knowledge materially handicaps MCP-driven behavior and
-search quality.
-
-### Viable Alternatives (Not Strict Zero-Knowledge)
-
-1. Encryption at rest/in transit with server-side decrypt in trusted runtime.
-   - Preserves MCP/search quality.
-   - Not zero-knowledge cryptographically.
-
-2. Client-side retrieval mode.
-   - Keep MCP/search local; cloud is sync/share/backup relay.
-   - Best for privacy-first users.
-   - Requires local agent availability for advanced retrieval.
-
-3. Limited encrypted indexing.
-   - Blind indexes for exact keywords only.
-   - No high-quality semantic search.
-   - Usually poor UX for natural-language memory recall.
-
-### Recommendation
-
-For Local+:
-
-1. Do not promise strict zero-knowledge for cloud MCP/search paths.
-2. Offer a privacy-first local mode where advanced retrieval stays local.
-3. Clearly label tradeoffs:
-   - "Local private mode" (best privacy, best local retrieval).
-   - "Cloud-assisted mode" (best cross-device/MCP consistency, trusted-runtime decrypt).
-
-This keeps messaging honest and avoids repeating the known incompatibility.
-

docs/specs/SPEC-SCHEMA-IMPL.md

@@ -1,368 +0,0 @@
-# SPEC-SCHEMA-IMPL: Schema System Implementation Plan
-
-**Status:** Draft
-**Created:** 2025-02-06
-**Branch:** `feature/schema-system`
-**Depends on:** [SPEC-SCHEMA](SPEC-SCHEMA.md)
-
-## Overview
-
-Implementation plan for the Basic Memory Schema System. The system is entirely programmatic —
-no LLM agent runtime or API key required. The LLM already in the user's session (Claude Code,
-Claude Desktop, etc.) provides the intelligence layer by reading schema notes via existing
-MCP tools.
-
-## Architecture
-
-```
-┌─────────────────────────────────────────────────┐
-│                   Entry Points                   │
-│  CLI (bm schema ...)  │  MCP (schema_validate)  │
-└──────────┬────────────┴──────────┬──────────────┘
-           │                       │
-           ▼                       ▼
-┌─────────────────────────────────────────────────┐
-│              Schema Service Layer                │
-│  resolve_schema · validate · infer · diff        │
-└──────────┬────────────────────────┬──────────────┘
-           │                       │
-           ▼                       ▼
-┌──────────────────────┐ ┌────────────────────────┐
-│   Picoschema Parser  │ │   Note/Entity Access   │
-│  YAML → SchemaModel  │ │  (existing repository) │
-└──────────────────────┘ └────────────────────────┘
-```
-
-No new database tables. Schemas are notes with `type: schema` — they're already indexed.
-Validation reads observations and relations from existing data.
-
-## Components
-
-### 1. Picoschema Parser
-
-**Location:** `src/basic_memory/schema/parser.py`
-
-Parses Picoschema YAML into an internal representation.
-
-```python
-@dataclass
-class SchemaField:
-    name: str
-    type: str                    # string, integer, number, boolean, any, or EntityName
-    required: bool               # True unless field name ends with ?
-    is_array: bool               # True if (array) notation
-    is_enum: bool                # True if (enum) notation
-    enum_values: list[str]       # Populated for enums
-    description: str | None      # Text after comma
-    is_entity_ref: bool          # True if type is capitalized (entity reference)
-    children: list[SchemaField]  # For (object) types
-
-
-@dataclass
-class SchemaDefinition:
-    entity: str                  # The entity type this schema describes
-    version: int                 # Schema version
-    fields: list[SchemaField]    # Parsed fields
-    validation_mode: str         # "warn" | "strict" | "off"
-    frontmatter_fields: list[SchemaField]  # From settings.frontmatter (default: [])
-
-
-def parse_picoschema(yaml_dict: dict) -> list[SchemaField]:
-    """Parse a Picoschema YAML dict into a list of SchemaField objects."""
-
-
-def parse_schema_note(frontmatter: dict) -> SchemaDefinition:
-    """Parse a full schema note's frontmatter into a SchemaDefinition."""
-```
-
-**Input/Output:**
-```yaml
-# Input (YAML dict from frontmatter)
-schema:
-  name: string, full name
-  role?: string, job title
-  works_at?: Organization, employer
-  expertise?(array): string, areas of knowledge
-```
-
-```python
-# Output
-[
-    SchemaField(name="name", type="string", required=True, description="full name", ...),
-    SchemaField(name="role", type="string", required=False, description="job title", ...),
-    SchemaField(name="works_at", type="Organization", required=False, is_entity_ref=True, ...),
-    SchemaField(name="expertise", type="string", required=False, is_array=True, ...),
-]
-```
-
-### 2. Schema Resolver
-
-**Location:** `src/basic_memory/schema/resolver.py`
-
-Finds the applicable schema for a note using the resolution order.
-
-```python
-async def resolve_schema(
-    note_frontmatter: dict,
-    search_fn: Callable,          # injected search capability
-) -> SchemaDefinition | None:
-    """Resolve schema for a note.
-
-    Resolution order:
-    1. Inline schema (frontmatter['schema'] is a dict)
-    2. Explicit reference (frontmatter['schema'] is a string)
-    3. Implicit by type (frontmatter['type'] → schema note with matching entity)
-    4. No schema (returns None)
-    """
-```
-
-### 3. Schema Validator
-
-**Location:** `src/basic_memory/schema/validator.py`
-
-Validates a note's observations and relations against a resolved schema.
-
-```python
-@dataclass
-class FieldResult:
-    field: SchemaField
-    status: str                  # "present" | "missing" | "type_mismatch"
-    values: list[str]            # Matched observation values or relation targets
-    message: str | None          # Human-readable detail
-
-
-@dataclass
-class ValidationResult:
-    note_identifier: str
-    schema_entity: str
-    passed: bool                 # True if no errors (warnings are OK)
-    field_results: list[FieldResult]
-    unmatched_observations: dict[str, int]   # category → count
-    unmatched_relations: list[str]           # relation types not in schema
-    warnings: list[str]
-    errors: list[str]
-
-
-async def validate_note(
-    note: Note,
-    schema: SchemaDefinition,
-    frontmatter: dict | None = None,
-) -> ValidationResult:
-    """Validate a note against a schema definition.
-
-    Mapping rules:
-    - field: string              → observation [field] exists
-    - field?(array): type        → multiple [field] observations
-    - field?: EntityType         → relation 'field [[...]]' exists
-    - field?(enum): [v]          → observation [field] value ∈ enum values
-    - settings.frontmatter field → frontmatter key presence/value
-    """
-```
-
-### 4. Schema Inference Engine
-
-**Location:** `src/basic_memory/schema/inference.py`
-
-Analyzes notes of a given type and suggests a schema based on usage frequency.
-
-```python
-@dataclass
-class FieldFrequency:
-    name: str
-    source: str                  # "observation" | "relation"
-    count: int                   # notes containing this field
-    total: int                   # total notes analyzed
-    percentage: float
-    sample_values: list[str]     # representative values
-    is_array: bool               # True if typically appears multiple times per note
-    target_type: str | None      # For relations, the most common target entity type
-
-
-@dataclass
-class InferenceResult:
-    entity_type: str
-    notes_analyzed: int
-    field_frequencies: list[FieldFrequency]
-    suggested_schema: dict       # Ready-to-use Picoschema YAML dict
-    suggested_required: list[str]
-    suggested_optional: list[str]
-    excluded: list[str]          # Below threshold
-
-
-async def infer_schema(
-    entity_type: str,
-    notes: list[Note],
-    required_threshold: float = 0.95,   # 95%+ = required
-    optional_threshold: float = 0.25,   # 25%+ = optional
-) -> InferenceResult:
-    """Analyze notes and suggest a Picoschema definition."""
-```
-
-### 5. Schema Diff
-
-**Location:** `src/basic_memory/schema/diff.py`
-
-Compares current note usage against an existing schema definition.
-
-```python
-@dataclass
-class SchemaDrift:
-    new_fields: list[FieldFrequency]       # Fields not in schema but common in notes
-    dropped_fields: list[FieldFrequency]   # Fields in schema but rare in notes
-    cardinality_changes: list[str]         # one → many or many → one
-    type_mismatches: list[str]             # observation values don't match declared type
-
-
-async def diff_schema(
-    schema: SchemaDefinition,
-    notes: list[Note],
-) -> SchemaDrift:
-    """Compare a schema against actual note usage to detect drift."""
-```
-
-## Entry Points
-
-### CLI Commands
-
-**Location:** `src/basic_memory/cli/schema.py`
-
-```python
-import typer
-
-schema_app = typer.Typer(name="schema", help="Schema management commands")
-
-@schema_app.command()
-async def validate(
-    target: str = typer.Argument(None, help="Note path or entity type"),
-    strict: bool = typer.Option(False, help="Override to strict mode"),
-):
-    """Validate notes against their schemas."""
-
-@schema_app.command()
-async def infer(
-    entity_type: str = typer.Argument(..., help="Entity type to analyze"),
-    threshold: float = typer.Option(0.25, help="Minimum frequency for optional fields"),
-    save: bool = typer.Option(False, help="Save to schema/ directory"),
-):
-    """Infer schema from existing notes of a type."""
-
-@schema_app.command()
-async def diff(
-    entity_type: str = typer.Argument(..., help="Entity type to diff"),
-):
-    """Show drift between schema and actual usage."""
-```
-
-Registered as subcommand: `bm schema validate`, `bm schema infer`, `bm schema diff`.
-
-### MCP Tools
-
-**Location:** `src/basic_memory/mcp/tools/schema.py`
-
-```python
-@mcp_tool
-async def schema_validate(
-    entity_type: str | None = None,
-    identifier: str | None = None,
-    project: str | None = None,
-) -> str:
-    """Validate notes against their resolved schema."""
-
-@mcp_tool
-async def schema_infer(
-    entity_type: str,
-    threshold: float = 0.25,
-    project: str | None = None,
-) -> str:
-    """Analyze existing notes and suggest a schema definition."""
-```
-
-### API Endpoints
-
-**Location:** `src/basic_memory/api/schema_router.py`
-
-```python
-router = APIRouter(prefix="/schema", tags=["schema"])
-
-@router.post("/validate")
-async def validate_schema(...) -> ValidationReport: ...
-
-@router.post("/infer")
-async def infer_schema(...) -> InferenceResult: ...
-
-@router.get("/diff/{entity_type}")
-async def diff_schema(...) -> SchemaDrift: ...
-```
-
-MCP tools call these endpoints via the typed client pattern (consistent with existing
-architecture).
-
-## Implementation Phases
-
-### Phase 1: Parser + Resolver
-
-Build the foundation — can parse Picoschema and find schemas for notes.
-
-**Deliverables:**
-- `schema/parser.py` — Picoschema YAML → `SchemaDefinition`
-- `schema/resolver.py` — Resolution order (inline → explicit ref → implicit by type → none)
-- Unit tests for all Picoschema syntax variations
-- Unit tests for resolution order
-
-**No external dependencies.** Pure Python parsing of YAML dicts. Can develop and test
-in isolation.
-
-### Phase 2: Validator
-
-Connect schemas to notes and produce validation results.
-
-**Deliverables:**
-- `schema/validator.py` — Validate note observations/relations against schema fields
-- API endpoint: `POST /schema/validate`
-- MCP tool: `schema_validate`
-- CLI command: `bm schema validate`
-- Integration tests with real notes and schemas
-
-**Depends on:** Phase 1 (parser + resolver)
-
-### Phase 3: Inference
-
-Analyze existing notes to suggest schemas.
-
-**Deliverables:**
-- `schema/inference.py` — Frequency analysis across notes of a type
-- API endpoint: `POST /schema/infer`
-- MCP tool: `schema_infer`
-- CLI command: `bm schema infer`
-- Option to save inferred schema as a note via `write_note`
-
-**Depends on:** Phase 1 (parser for output format)
-
-### Phase 4: Diff
-
-Compare schemas against current usage.
-
-**Deliverables:**
-- `schema/diff.py` — Drift detection between schema and actual notes
-- API endpoint: `GET /schema/diff/{entity_type}`
-- CLI command: `bm schema diff`
-
-**Depends on:** Phase 1 (parser), Phase 3 (inference, for frequency analysis)
-
-## Testing Strategy
-
-- **Unit tests** (`tests/schema/`): Parser edge cases, resolution logic, validation mapping,
-  inference thresholds
-- **Integration tests** (`test-int/schema/`): End-to-end with real markdown files, schema notes
-  on disk, CLI invocation
-- Coverage target: 100% (consistent with project standard)
-
-## What This Does NOT Include
-
-- No new database tables or migrations
-- No new markdown syntax (schemas validate existing observations/relations)
-- No LLM agent runtime or API key management
-- No hook integration (deferred)
-- No schema composition/inheritance (deferred)
-- No OWL/RDF export (deferred)
-- No built-in templates (deferred)

docs/specs/SPEC-SCHEMA.md

@@ -1,492 +0,0 @@
-# SPEC-SCHEMA: Basic Memory Schema System
-
-**Status:** Draft
-**Created:** 2025-02-06
-**Branch:** `feature/schema-system`
-
-## Summary
-
-A schema system for Basic Memory that uses [Picoschema](https://genkit.dev/docs/dotprompt/)
-syntax in YAML frontmatter. Schemas validate notes against their existing observation/relation
-structure — no new data model, no migration, just a declarative lens over what's already there.
-
-## Core Principles
-
-1. **Schemas are just notes** — A schema is a note with `type: schema`, lives anywhere
-2. **Use prior art** — Picoschema syntax in YAML frontmatter, no custom notation
-3. **Validation maps to existing format** — Observations and relations, not a parallel data model
-4. **Validation is soft** — Warnings by default, not blocking errors
-5. **Inference over prescription** — Schemas describe reality, emerge from usage
-6. **No built-in agent** — Programmatic core; the LLM already in the session provides intelligence
-
-## Picoschema Syntax
-
-Picoschema is a compact schema notation from Google's Dotprompt that fits naturally in YAML
-frontmatter.
-
-### Supported Types
-
-| Type | Description |
-|------|-------------|
-| `string` | Text value |
-| `integer` | Whole number |
-| `number` | Decimal number |
-| `boolean` | True/false |
-| `any` | Any scalar type |
-| `EntityName` | Reference to another entity (capitalized = entity reference) |
-
-### Syntax Rules
-
-```yaml
-schema:
-  name: string, full name                # required field with description
-  email?: string, contact email           # ? = optional
-  role?: string, job title
-  works_at?: Organization, employer       # capitalized type = entity reference
-  tags?(array): string, categories        # array of type
-  status?(enum): [active, inactive]       # enum with allowed values
-  metadata?(object):                      # nested object
-    updated_at?: string
-    source?: string
-```
-
-- `field: type` — required field
-- `field?: type` — optional field
-- `field(array): type` — array of values
-- `field?(enum): [values]` — enumeration
-- `field?(object):` — nested object with sub-fields
-- `, description` — description after comma
-- `EntityName` as type (capitalized) — reference to another entity
-
-## Schema-to-Note Mapping
-
-Schemas validate against the existing Basic Memory note format. No new syntax for note
-authors to learn.
-
-### Mapping Rules
-
-| Schema Declaration | Grounded In | Example Match |
-|--------------------|-------------|---------------|
-| `field: string` | Observation `[field] value` | `- [name] Paul Graham` |
-| `field?(array): string` | Multiple `[field]` observations | `- [expertise] Lisp` (×N) |
-| `field?: EntityType` | Relation `field [[Target]]` | `- works_at [[Y Combinator]]` |
-| `field?(array): EntityType` | Multiple `field` relations | `- authored [[Book]]` (×N) |
-| `tags` | Frontmatter `tags` array | `tags: [startups, essays]` |
-| `field?(enum): [values]` | Observation `[field] value` where value ∈ set | `- [status] active` |
-| `settings.frontmatter` field | Frontmatter key presence/value | `tags: [python, ai]` |
-
-### Key Insight
-
-Schemas don't introduce a new way to store data. They describe the patterns already present
-in observations and relations. A note doesn't have to change how it's written — the schema
-just says "a good Person note has a `[name]` observation and a `works_at` relation."
-
-## Schema Definition
-
-### As a Dedicated Schema Note
-
-```yaml
-# schema/Person.md
----
-title: Person
-type: schema
-entity: Person
-version: 1
-schema:
-  name: string, full name
-  email?: string, contact email
-  role?: string, job title
-  works_at?: Organization, employer
-  expertise?(array): string, areas of knowledge
-settings:
-  validation: warn    # warn | strict | off
-  frontmatter:
-    tags?(array): string, note categories
-    status?(enum): [draft, review, published]
----
-
-# Person
-
-A human individual in the knowledge graph.
-
-Any documentation about this entity type goes here as prose.
-```
-
-Schema notes are regular Basic Memory notes. They show up in search, can have their own
-observations and relations, and can be organized in any folder (though `schema/` is
-the suggested convention).
-
-### Inline Schema in a Note
-
-Notes can carry their own schema directly:
-
-```yaml
-# meetings/2024-01-15-standup.md
----
-title: Team Standup 2024-01-15
-type: meeting
-schema:
-  attendees(array): string, who was there
-  decisions(array): string, what was decided
-  action_items(array): string, follow-ups
-  blockers?(array): string, anything stuck
----
-
-# Team Standup 2024-01-15
-
-## Observations
-- [attendees] Paul
-- [attendees] Sarah
-- [decisions] Ship v2 by Friday
-- [action_items] Paul to review PR #42
-- [blockers] Waiting on API credentials
-```
-
-Good for one-off structured notes or prototyping a schema before extracting it.
-
-### Explicit Schema Reference
-
-A note can reference a schema by entity name or permalink:
-
-```yaml
-# projects/basic-memory.md
----
-title: Basic Memory
-schema: SoftwareProject         # by entity name
----
-
-# research/llm-memory-patterns.md
----
-title: LLM Memory Patterns
-schema: schema/research-project  # by permalink
----
-```
-
-Use cases:
-- Note's `type` differs from the schema it should validate against
-- Multiple schema variants exist for the same domain
-- Applying structure to existing notes without changing their type
-
-## Schema Resolution
-
-When validating a note, schemas resolve in priority order:
-
-```
-1. Inline schema     →  schema: { ... }        (dict in frontmatter)
-2. Explicit ref      →  schema: Person          (string in frontmatter)
-3. Implicit by type  →  type: Person            (lookup schema note with entity: Person)
-4. No schema         →  no validation           (perfectly fine)
-```
-
-```python
-async def resolve_schema(note: Note) -> Schema | None:
-    schema_value = note.frontmatter.get('schema')
-
-    # 1. Inline schema (dict)
-    if isinstance(schema_value, dict):
-        return parse_picoschema(schema_value)
-
-    # 2. Explicit reference (string)
-    if isinstance(schema_value, str):
-        schema_note = await find_schema_note(schema_value)
-        if schema_note:
-            return parse_picoschema(schema_note.frontmatter['schema'])
-
-    # 3. Implicit by type
-    note_type = note.frontmatter.get('type')
-    if note_type:
-        results = await search_notes(f"type:schema entity:{note_type}")
-        if results:
-            return parse_picoschema(results[0].frontmatter['schema'])
-
-    # 4. No schema
-    return None
-```
-
-## Validation
-
-### Modes
-
-Configured in the schema's `settings.validation`:
-
-| Mode | Behavior |
-|------|----------|
-| `off` | No validation |
-| `warn` | Warnings in output, doesn't block (default) |
-| `strict` | Errors that block sync, for CI/CD enforcement |
-
-### Validation Output
-
-For a note missing required fields:
-
-```
-$ bm schema validate people/ada-lovelace.md
-
-⚠ Person schema validation:
-  - Missing required field: name (expected [name] observation)
-  - Missing optional field: role
-  - Missing optional field: works_at (no relation found)
-
-ℹ Unmatched observations: [fact] ×2, [born] ×1
-ℹ Unmatched relations: collaborated_with
-```
-
-"Unmatched" items are informational — observations and relations the schema doesn't cover.
-They're valid. Schemas are a subset, not a straitjacket.
-
-### Frontmatter Validation
-
-Schema notes can declare validation rules for frontmatter keys under `settings.frontmatter`
-using the same Picoschema syntax as the `schema` block:
-
-```yaml
-settings:
-  validation: warn
-  frontmatter:
-    tags?(array): string
-    status?(enum): [draft, review, published]
-```
-
-- Frontmatter rules use the same Picoschema key syntax (`?` for optional, `(enum)`, `(array)`)
-- Only available on schema notes (inline schemas skip frontmatter validation)
-- Checks key presence (required vs optional) and enum value membership
-- Unmatched frontmatter keys not in the schema are silently ignored
-- Missing required frontmatter keys produce a warning (or error in strict mode)
-
-Example output for a missing required frontmatter key:
-
-```
-⚠ Person schema validation:
-  - Missing required frontmatter key: status
-```
-
-### Batch Validation
-
-```
-$ bm schema validate Person
-
-Validating 30 notes against Person schema...
-
-✓ people/paul-graham.md         — all fields present
-✓ people/rich-hickey.md         — all fields present
-⚠ people/ada-lovelace.md        — missing: name
-⚠ people/alan-kay.md            — missing: name, role
-✓ people/linus-torvalds.md      — all fields present
-...
-
-Summary: 22/30 valid, 8 warnings, 0 errors
-```
-
-## Emerging Schemas
-
-### The Problem with Traditional Schemas
-
-Most schema systems require: define schema → create conforming content → fight the schema
-when reality doesn't match. This is backwards. Knowledge grows organically.
-
-### The Basic Memory Approach
-
-```
-Write notes freely → Patterns emerge → Crystallize into schema → Validate future notes
-```
-
-### Schema Inference
-
-Generate schemas from existing notes by analyzing observation and relation frequency:
-
-```
-$ bm schema infer Person
-
-Analyzing 30 notes with type: Person...
-
-Observations found:
-  [name]        30/30  100%  → name: string
-  [role]        27/30   90%  → role?: string
-  [fact]        25/30   83%  (generic — no single field)
-  [expertise]   18/30   60%  → expertise?(array): string
-  [email]        8/30   27%  → email?: string
-  [born]         6/30   20%  (below threshold)
-
-Relations found:
-  works_at      22/30   73%  → works_at?: Organization
-  authored      11/30   37%  → authored?(array): string
-
-Suggested schema:
-  name: string, full name
-  role?: string, job title
-  expertise?(array): string, areas of knowledge
-  email?: string, contact email
-  works_at?: Organization, employer
-
-Save to schema/Person.md? [y/n]
-```
-
-Frequency thresholds:
-- 100% present → required field
-- 25%+ present → optional field
-- Below 25% → excluded from suggestion (but noted)
-
-### Schema Drift Detection
-
-Track how usage patterns shift over time:
-
-```
-$ bm schema diff Person
-
-Schema drift detected:
-
-+ expertise: now in 81% of notes (was 12%)
-- department: dropped to 3% of notes
-~ works_at: cardinality changed (one → many)
-
-Update schema? [y/n/review]
-```
-
-## LLM Integration (AI Guidance)
-
-No agent runtime or API key required. The LLM already in the session uses schemas as
-context for note creation.
-
-### Flow
-
-1. User asks LLM to "write a note about Rich Hickey"
-2. LLM determines `type: Person` is appropriate
-3. LLM calls `search_notes("type:schema entity:Person")` → finds schema
-4. LLM reads schema fields: required `name`, optional `role`, `works_at`, `expertise`
-5. LLM calls `write_note` with observations and relations that satisfy the schema
-
-The schema acts as a creation template. The LLM knows what a "complete" note looks like
-without any custom agent infrastructure.
-
-### MCP Tools
-
-```python
-@mcp_tool
-async def schema_validate(
-    entity_type: str | None = None,
-    identifier: str | None = None,
-    project: str | None = None,
-) -> ValidationReport:
-    """Validate notes against their resolved schema.
-
-    Validates a specific note (by identifier) or all notes of a given type.
-    Returns warnings/errors based on the schema's validation mode.
-    """
-
-@mcp_tool
-async def schema_infer(
-    entity_type: str,
-    threshold: float = 0.25,
-    project: str | None = None,
-) -> SuggestedSchema:
-    """Analyze existing notes and suggest a schema definition.
-
-    Examines observation categories and relation types across all notes
-    of the given type. Returns frequency analysis and suggested Picoschema.
-    """
-```
-
-## CLI Commands
-
-```bash
-# Validate a specific note
-bm schema validate people/ada-lovelace.md
-
-# Validate all notes of a type
-bm schema validate Person
-
-# Validate everything with a schema
-bm schema validate
-
-# Infer schema from existing notes
-bm schema infer Person
-
-# Show schema drift from current definition
-bm schema diff Person
-
-# List all schema notes
-bm search "type:schema"
-```
-
-## Examples
-
-### Complete Person Workflow
-
-**Schema:**
-```yaml
-# schema/Person.md
----
-title: Person
-type: schema
-entity: Person
-version: 1
-schema:
-  name: string, full name
-  role?: string, job title or position
-  works_at?: Organization, employer
-  expertise?(array): string, areas of knowledge
-  email?: string, contact email
-settings:
-  validation: warn
----
-
-# Person
-
-A human individual in the knowledge graph.
-```
-
-**Valid note:**
-```yaml
-# people/paul-graham.md
----
-title: Paul Graham
-type: Person
-tags: [startups, essays, lisp]
----
-
-# Paul Graham
-
-## Observations
-- [name] Paul Graham
-- [role] Essayist and investor
-- [expertise] Startups
-- [expertise] Lisp
-- [expertise] Essay writing
-- [fact] Created Viaweb, the first web app
-
-## Relations
-- works_at [[Y Combinator]]
-- authored [[Hackers and Painters]]
-```
-
-**Note with warnings:**
-```yaml
-# people/ada-lovelace.md
----
-title: Ada Lovelace
-type: Person
----
-
-# Ada Lovelace
-
-## Observations
-- [fact] Wrote the first computer program
-- [born] 1815
-
-## Relations
-- collaborated_with [[Charles Babbage]]
-```
-
-Validation: warns about missing required `[name]` observation. Everything else is optional
-or unmatched (which is fine).
-
-## Future Considerations (Deferred)
-
-These are interesting but out of scope for the initial implementation:
-
-- **Multiple schema inheritance** — `schema: [Person, Author]`
-- **Hook integration** — Pre-write validation via the hooks system
-- **OWL/RDF export** — `bm schema export --format owl`
-- **SPARQL queries** — Schema-aware graph queries
-- **Built-in templates** — `bm schema use gtd`, `bm schema use zettelkasten`
-- **Schema versioning/migration** — Tracking breaking changes across versions

docs/testing-coverage.md

@@ -1,28 +0,0 @@
-## Coverage policy (practical 100%)
-
-Basic Memory’s test suite intentionally mixes:
-- unit tests (fast, deterministic)
-- integration tests (real filesystem + real DB via `test-int/`)
-
-To keep the default CI signal **stable and meaningful**, the default `pytest` coverage report targets **core library logic** and **excludes** a small set of modules that are either:
-- highly environment-dependent (OS/DB tuning)
-- inherently interactive (CLI)
-- background-task orchestration (watchers/sync runners)
-
-### What's excluded (and why)
-
-Coverage excludes are configured in `pyproject.toml` under `[tool.coverage.report].omit`.
-
-Current exclusions include:
-- `src/basic_memory/cli/**`: interactive wrappers; behavior is validated via higher-level tests and smoke tests.
-- `src/basic_memory/db.py`: platform/backend tuning paths (SQLite/Postgres/Windows), covered by integration tests and targeted runs.
-- `src/basic_memory/services/initialization.py`: startup orchestration/background tasks; covered indirectly by app/MCP entrypoints.
-- `src/basic_memory/sync/sync_service.py`: heavy filesystem↔DB integration; validated in integration suite (not enforced in unit coverage).
-
-### Recommended additional runs
-
-If you want extra confidence locally/CI:
-- **Postgres backend**: run tests with `BASIC_MEMORY_TEST_POSTGRES=1`.
-- **Strict backend-complete coverage**: run coverage on SQLite + Postgres and combine the results (recommended).
-
-

justfile

@@ -2,182 +2,27 @@
 
 # Install dependencies
 install:
-    uv sync
-    @echo ""
-    @echo "💡 Remember to activate the virtual environment by running: source .venv/bin/activate"
+    pip install -e ".[dev]"
 
-# ==============================================================================
-# DATABASE BACKEND TESTING
-# ==============================================================================
-# Basic Memory supports dual database backends (SQLite and Postgres).
-# By default, tests run against SQLite (fast, no dependencies).
-# Set BASIC_MEMORY_TEST_POSTGRES=1 to run against Postgres (uses testcontainers).
-#
-# Quick Start:
-#   just test              # Run all tests against SQLite (default)
-#   just test-sqlite       # Run all tests against SQLite
-#   just test-postgres     # Run all tests against Postgres (testcontainers)
-#   just test-unit-sqlite  # Run unit tests against SQLite
-#   just test-unit-postgres # Run unit tests against Postgres
-#   just test-int-sqlite   # Run integration tests against SQLite
-#   just test-int-postgres # Run integration tests against Postgres
-#
-# CI runs both in parallel for faster feedback.
-# ==============================================================================
+# Run unit tests in parallel
+test-unit:
+    uv run pytest -p pytest_mock -v -n auto
 
-# Run all tests against SQLite and Postgres
-test: test-sqlite test-postgres
+# Run integration tests in parallel
+test-int:
+    uv run pytest -p pytest_mock -v --no-cov -n auto test-int
 
-# Run all tests against SQLite
-test-sqlite: test-unit-sqlite test-int-sqlite
-
-# Run all tests against Postgres (uses testcontainers)
-test-postgres: test-unit-postgres test-int-postgres
-
-# Run unit tests against SQLite
-test-unit-sqlite:
-    BASIC_MEMORY_ENV=test uv run pytest -p pytest_mock -v --no-cov tests
-
-# Run unit tests against Postgres
-test-unit-postgres:
-    BASIC_MEMORY_ENV=test BASIC_MEMORY_TEST_POSTGRES=1 uv run pytest -p pytest_mock -v --no-cov tests
-
-# Run integration tests against SQLite (excludes semantic benchmarks — use just test-semantic)
-test-int-sqlite:
-    BASIC_MEMORY_ENV=test uv run pytest -p pytest_mock -v --no-cov -m "not semantic" test-int
-
-# Run integration tests against Postgres
-# Note: Uses timeout due to FastMCP Client + asyncpg cleanup hang (tests pass, process hangs on exit)
-# See: https://github.com/jlowin/fastmcp/issues/1311
-test-int-postgres:
-    #!/usr/bin/env bash
-    set -euo pipefail
-    # Use gtimeout (macOS/Homebrew) or timeout (Linux)
-    TIMEOUT_CMD=$(command -v gtimeout || command -v timeout || echo "")
-    if [[ -n "$TIMEOUT_CMD" ]]; then
-        $TIMEOUT_CMD --signal=KILL 600 bash -c 'BASIC_MEMORY_ENV=test BASIC_MEMORY_TEST_POSTGRES=1 uv run pytest -p pytest_mock -v --no-cov -m "not semantic" test-int' || test $? -eq 137
-    else
-        echo "⚠️  No timeout command found, running without timeout..."
-        BASIC_MEMORY_ENV=test BASIC_MEMORY_TEST_POSTGRES=1 uv run pytest -p pytest_mock -v --no-cov -m "not semantic" test-int
-    fi
-
-# Run tests impacted by recent changes (requires pytest-testmon)
-testmon *args:
-    BASIC_MEMORY_ENV=test uv run pytest -p pytest_mock -v --no-cov --testmon --testmon-forceselect {{args}}
-
-# Run MCP smoke test (fast end-to-end loop)
-test-smoke:
-    BASIC_MEMORY_ENV=test uv run pytest -p pytest_mock -v --no-cov -m smoke test-int/mcp/test_smoke_integration.py
-
-# Fast local loop: lint, format, typecheck, impacted tests
-fast-check:
-    just fix
-    just format
-    just typecheck
-    just testmon
-    just test-smoke
-
-# Reset Postgres test database (drops and recreates schema)
-# Useful when Alembic migration state gets out of sync during development
-# Uses credentials from docker-compose-postgres.yml
-postgres-reset:
-    docker exec basic-memory-postgres psql -U ${POSTGRES_USER:-basic_memory_user} -d ${POSTGRES_TEST_DB:-basic_memory_test} -c "DROP SCHEMA public CASCADE; CREATE SCHEMA public;"
-    @echo "✅ Postgres test database reset"
-
-# Run Alembic migrations manually against Postgres test database
-# Useful for debugging migration issues
-# Uses credentials from docker-compose-postgres.yml (can override with env vars)
-postgres-migrate:
-    @cd src/basic_memory/alembic && \
-    BASIC_MEMORY_DATABASE_BACKEND=postgres \
-    BASIC_MEMORY_DATABASE_URL=${POSTGRES_TEST_URL:-postgresql+asyncpg://basic_memory_user:dev_password@localhost:5433/basic_memory_test} \
-    uv run alembic upgrade head
-    @echo "✅ Migrations applied to Postgres test database"
-
-# Run Windows-specific tests only (only works on Windows platform)
-# These tests verify Windows-specific database optimizations (locking mode, NullPool)
-# Will be skipped automatically on non-Windows platforms
-test-windows:
-    BASIC_MEMORY_ENV=test uv run pytest -p pytest_mock -v --no-cov -m windows tests test-int
-
-# Run benchmark tests only (performance testing)
-# These are slow tests that measure sync performance with various file counts
-# Excluded from default test runs to keep CI fast
-test-benchmark:
-    BASIC_MEMORY_ENV=test uv run pytest -p pytest_mock -v --no-cov -m benchmark tests test-int
-
-# Run semantic search quality benchmarks (all combos)
-test-semantic:
-    BASIC_MEMORY_ENV=test uv run pytest -p pytest_mock -v --no-cov -m semantic test-int/semantic/
-
-# Run semantic benchmarks with JSON artifact output, then show report
-test-semantic-report:
-    BASIC_MEMORY_ENV=test BASIC_MEMORY_BENCHMARK_OUTPUT=.benchmarks/semantic-quality.jsonl uv run pytest -p pytest_mock -v -s --no-cov -m semantic test-int/semantic/
-    uv run python test-int/semantic/report.py .benchmarks/semantic-quality.jsonl
-
-# Run semantic benchmarks (Postgres combos only)
-test-semantic-postgres:
-    BASIC_MEMORY_ENV=test uv run pytest -p pytest_mock -v --no-cov -m semantic -k postgres test-int/semantic/
-
-# View semantic benchmark results (rich formatted table)
-# Usage: just semantic-report [--filter-combo sqlite] [--filter-suite paraphrase] [--sort-by avg_latency_ms]
-semantic-report *args:
-    uv run python test-int/semantic/report.py .benchmarks/semantic-quality.jsonl {{args}}
-
-# Compare two search benchmark JSONL outputs
-# Usage:
-#   just benchmark-compare .benchmarks/search-baseline.jsonl .benchmarks/search-candidate.jsonl
-#   just benchmark-compare .benchmarks/search-baseline.jsonl .benchmarks/search-candidate.jsonl --format markdown --show-missing
-benchmark-compare baseline candidate *args:
-    uv run python test-int/compare_search_benchmarks.py "{{baseline}}" "{{candidate}}" --format table {{args}}
-
-# Run all tests including Windows, Postgres, and Benchmarks (for CI/comprehensive testing)
-# Use this before releasing to ensure everything works across all backends and platforms
-test-all:
-    BASIC_MEMORY_ENV=test uv run pytest -p pytest_mock -v --no-cov tests test-int
-
-# Generate HTML coverage report
-coverage:
-    #!/usr/bin/env bash
-    set -euo pipefail
-    
-    uv run coverage erase
-    
-    echo "🔎 Coverage (SQLite)..."
-    BASIC_MEMORY_ENV=test uv run coverage run --source=basic_memory -m pytest -p pytest_mock -v --no-cov tests test-int
-    
-    echo "🔎 Coverage (Postgres via testcontainers)..."
-    # Note: Uses timeout due to FastMCP Client + asyncpg cleanup hang (tests pass, process hangs on exit)
-    # See: https://github.com/jlowin/fastmcp/issues/1311
-    TIMEOUT_CMD=$(command -v gtimeout || command -v timeout || echo "")
-    if [[ -n "$TIMEOUT_CMD" ]]; then
-        $TIMEOUT_CMD --signal=KILL 600 bash -c 'BASIC_MEMORY_ENV=test BASIC_MEMORY_TEST_POSTGRES=1 uv run coverage run --source=basic_memory -m pytest -p pytest_mock -v --no-cov -m postgres tests test-int' || test $? -eq 137
-    else
-        echo "⚠️  No timeout command found, running without timeout..."
-        BASIC_MEMORY_ENV=test BASIC_MEMORY_TEST_POSTGRES=1 uv run coverage run --source=basic_memory -m pytest -p pytest_mock -v --no-cov -m postgres tests test-int
-    fi
-    
-    echo "🧩 Combining coverage data..."
-    uv run coverage combine
-    uv run coverage report -m
-    uv run coverage html
-    echo "Coverage report generated in htmlcov/index.html"
-
-# Lint and fix code (calls fix)
-lint: fix
+# Run all tests
+test: test-unit test-int
 
 # Lint and fix code
-fix:
-    uv run ruff check --fix --unsafe-fixes src tests test-int
+lint:
+    ruff check . --fix
 
-# Type check code (pyright)
-typecheck:
+# Type check code
+type-check:
     uv run pyright
 
-# Type check code (ty)
-typecheck-ty:
-    uv run ty check src/
-
 # Clean build artifacts and cache files
 clean:
     find . -type f -name '*.pyc' -delete
@@ -193,28 +38,21 @@ format:
 run-inspector:
     npx @modelcontextprotocol/inspector
 
-# Run doctor checks in an isolated temp home/config
-doctor:
-    #!/usr/bin/env bash
-    set -euo pipefail
-    TMP_HOME=$(mktemp -d)
-    TMP_CONFIG=$(mktemp -d)
-    HOME="$TMP_HOME" \
-    BASIC_MEMORY_ENV=test \
-    BASIC_MEMORY_HOME="$TMP_HOME/basic-memory" \
-    BASIC_MEMORY_CONFIG_DIR="$TMP_CONFIG" \
-    ./.venv/bin/python -m basic_memory.cli.main doctor --local
+# Build macOS installer
+installer-mac:
+    cd installer && chmod +x make_icons.sh && ./make_icons.sh
+    cd installer && uv run python setup.py bdist_mac
 
+# Build Windows installer
+installer-win:
+    cd installer && uv run python setup.py bdist_win32
 
 # Update all dependencies to latest versions
 update-deps:
     uv sync --upgrade
 
 # Run all code quality checks and tests
-check: lint format typecheck test
-
-# Run all code quality checks and all test suites, including semantic benchmarks
-check-all: lint format typecheck test test-semantic
+check: lint format type-check test
 
 # Generate Alembic migration with descriptive message
 migration message:
@@ -255,22 +93,16 @@ release version:
     fi
     
     # Run quality checks
-    echo "🔍 Running lint  checks..."
-    just lint
-    just typecheck
+    echo "🔍 Running quality checks..."
+    just check
     
     # Update version in __init__.py
     echo "📝 Updating version in __init__.py..."
     sed -i.bak "s/__version__ = \".*\"/__version__ = \"$VERSION_NUM\"/" src/basic_memory/__init__.py
     rm -f src/basic_memory/__init__.py.bak
-
-    # Update version in server.json (MCP registry metadata)
-    echo "📝 Updating version in server.json..."
-    sed -i.bak "s/\"version\": \"[^\"]*\"/\"version\": \"$VERSION_NUM\"/g" server.json
-    rm -f server.json.bak
-
+    
     # Commit version update
-    git add src/basic_memory/__init__.py server.json
+    git add src/basic_memory/__init__.py
     git commit -m "chore: update version to $VERSION_NUM for {{version}} release"
     
     # Create and push tag
@@ -284,12 +116,6 @@ release version:
     echo "✅ Release {{version}} created successfully!"
     echo "📦 GitHub Actions will build and publish to PyPI"
     echo "🔗 Monitor at: https://github.com/basicmachines-co/basic-memory/actions"
-    echo ""
-    echo "📝 REMINDER: Post-release tasks:"
-    echo "   1. docs.basicmemory.com - Add release notes to src/pages/latest-releases.mdx"
-    echo "   2. basicmachines.co - Update version in src/components/sections/hero.tsx"
-    echo "   3. MCP Registry - Run: mcp-publisher publish"
-    echo "   See: .claude/commands/release/release.md for detailed instructions"
 
 # Create a beta release (e.g., just beta v0.13.2b1)
 beta version:
@@ -326,22 +152,16 @@ beta version:
     fi
     
     # Run quality checks
-    echo "🔍 Running lint  checks..."
-    just lint
-    just typecheck
+    echo "🔍 Running quality checks..."
+    just check
     
     # Update version in __init__.py
     echo "📝 Updating version in __init__.py..."
     sed -i.bak "s/__version__ = \".*\"/__version__ = \"$VERSION_NUM\"/" src/basic_memory/__init__.py
     rm -f src/basic_memory/__init__.py.bak
-
-    # Update version in server.json (MCP registry metadata)
-    echo "📝 Updating version in server.json..."
-    sed -i.bak "s/\"version\": \"[^\"]*\"/\"version\": \"$VERSION_NUM\"/g" server.json
-    rm -f server.json.bak
-
+    
     # Commit version update
-    git add src/basic_memory/__init__.py server.json
+    git add src/basic_memory/__init__.py
     git commit -m "chore: update version to $VERSION_NUM for {{version}} beta release"
     
     # Create and push tag
@@ -356,12 +176,41 @@ beta version:
     echo "📦 GitHub Actions will build and publish to PyPI as pre-release"
     echo "🔗 Monitor at: https://github.com/basicmachines-co/basic-memory/actions"
     echo "📥 Install with: uv tool install basic-memory --pre"
-    echo ""
-    echo "📝 REMINDER: For stable releases, update documentation sites:"
-    echo "   1. docs.basicmemory.com - Add release notes to src/pages/latest-releases.mdx"
-    echo "   2. basicmachines.co - Update version in src/components/sections/hero.tsx"
-    echo "   See: .claude/commands/release/release.md for detailed instructions"
+
+# Build DXT package with bundled virtual environment
+dxt:
+    #!/usr/bin/env bash
+    echo "🏗️  Building DXT package..."
+    
+    # Clean and create bundle directory
+    rm -rf dxt-bundle
+    mkdir -p dxt-bundle/server/lib dxt-bundle/src
+    
+    # Bundle dependencies to server/lib (like Anthropic example)
+    echo "📦 Bundling Python dependencies..."
+    uv pip install -e . --target dxt-bundle/server/lib --force-reinstall
+    
+    # Copy source code
+    echo "📄 Copying source code..."
+    cp -r src/basic_memory dxt-bundle/src/
+    
+    # Copy manifest
+    cp manifest.json dxt-bundle/
+    
+    # Copy assets (icons, screenshots)
+    echo "🖼️  Copying assets..."
+    mkdir -p dxt-bundle/images
+    cp images/disk-ai-logo-black-fg-white-bg.png dxt-bundle/images/
+    cp images/claude-*.png dxt-bundle/images/
+    
+    # Create DXT package
+    echo "📦 Creating DXT package..."
+    cd dxt-bundle && dxt pack . ../basic-memory.dxt
+    
+    echo "✅ DXT package created: basic-memory.dxt"
+    echo "🔍 Info: dxt info basic-memory.dxt"
+    echo "🧪 Test: Install in Claude Desktop"
 
 # List all available recipes
 default:
-    @just --list
+    @just --list

llms-install.md

@@ -1,144 +0,0 @@
-# Basic Memory Installation Guide for LLMs
-
-This guide is specifically designed to help AI assistants like Cline install and configure Basic Memory. Follow these
-steps in order.
-
-## Installation Steps
-
-### 1. Install Basic Memory Package
-
-Use one of the following package managers to install:
-
-```bash
-# Install with uv (recommended)
-uv tool install basic-memory
-
-# Or with pip
-pip install basic-memory
-```
-
-### 2. Configure MCP Server
-
-Add the following to your config:
-
-```json
-{
-  "mcpServers": {
-    "basic-memory": {
-      "command": "uvx",
-      "args": [
-        "basic-memory",
-        "mcp"
-      ]
-    }
-  }
-}
-```
-
-For Claude Desktop, this file is located at:
-
-macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
-Windows: %APPDATA%\Claude\claude_desktop_config.json
-
-### 3. Start Synchronization (optional)
-
-To synchronize files in real-time, run:
-
-```bash
-basic-memory sync --watch
-```
-
-Or for a one-time sync:
-
-```bash
-basic-memory sync
-```
-
-### 4. Updating Basic Memory
-
-Basic Memory supports automatic updates by default for `uv tool` and Homebrew installs.
-
-For manual checks and upgrades:
-
-```bash
-# Check now and install if supported
-bm update
-
-# Check only, do not install
-bm update --check
-```
-
-To disable automatic updates, set `"auto_update": false` in `~/.basic-memory/config.json`.
-
-## Configuration Options
-
-### Custom Directory
-
-To use a directory other than the default `~/basic-memory`:
-
-```bash
-basic-memory project add custom-project /path/to/your/directory
-basic-memory project default custom-project
-```
-
-### Multiple Projects
-
-To manage multiple knowledge bases:
-
-```bash
-# List all projects
-basic-memory project list
-
-# Add a new project
-basic-memory project add work ~/work-basic-memory
-
-# Set default project
-basic-memory project default work
-```
-
-## Importing Existing Data
-
-### From Claude.ai
-
-```bash
-basic-memory import claude conversations path/to/conversations.json
-basic-memory import claude projects path/to/projects.json
-```
-
-### From ChatGPT
-
-```bash
-basic-memory import chatgpt path/to/conversations.json
-```
-
-### From MCP Memory Server
-
-```bash
-basic-memory import memory-json path/to/memory.json
-```
-
-## Troubleshooting
-
-If you encounter issues:
-
-1. Check that Basic Memory is properly installed:
-   ```bash
-   basic-memory --version
-   ```
-
-2. Verify the sync process is running:
-   ```bash
-   ps aux | grep basic-memory
-   ```
-
-3. Check sync output for errors:
-   ```bash
-   basic-memory sync --verbose
-   ```
-
-4. Check log output:
-   ```bash
-   cat ~/.basic-memory/basic-memory.log
-   ```
-
-For more detailed information, refer to the [full documentation](https://docs.basicmemory.com/).

manifest.json

@@ -0,0 +1,166 @@
+{
+  "dxt_version": "0.1",
+  "name": "basic-memory",
+  "display_name": "Basic Memory Knowledge Management",
+  "version": "0.14.0",
+  "description": "Local-first knowledge management with AI integration",
+  "long_description": "Basic Memory is a local-first knowledge management system that creates a personal knowledge graph from markdown files. It enables bidirectional communication between LLMs and your local knowledge base through the Model Context Protocol (MCP), allowing AI assistants to read, write, and navigate your knowledge seamlessly.",
+  "author": {
+    "name": "Basic Machines",
+    "email": "hello@basicmachines.co",
+    "url": "https://github.com/basicmachines-co"
+  },
+  "homepage": "https://basicmemory.com/",
+  "documentation": "https://memory.basicmachines.co/",
+  "support": "https://memory.basicmachines.co/user-guide",
+  "icon": "images/disk-ai-logo-black-fg-white-bg.png",
+  "screenshots": [
+    "images/claude-continue-conversation.png",
+    "images/claude-switch-project.png",
+    "images/claude-view-note.png"
+  ],
+  "server": {
+    "type": "python",
+    "entry_point": "src/basic_memory/cli/main.py",
+    "mcp_config": {
+      "command": "python3",
+      "args": [
+        "${__dirname}/src/basic_memory/cli/main.py",
+        "mcp",
+        "--transport=stdio"
+      ],
+      "env": {
+        "BASIC_MEMORY_HOME": "${user_data_dir}/basic-memory",
+        "PYTHONPATH": "${__dirname}/server/lib:${__dirname}/src"
+      }
+    }
+  },
+  "tools": [
+    {
+      "name": "write_note",
+      "description": "Create or update markdown notes with semantic structure"
+    },
+    {
+      "name": "read_note",
+      "description": "Read notes by title, permalink, or memory:// URL"
+    },
+    {
+      "name": "edit_note",
+      "description": "Edit notes incrementally (append, prepend, find/replace)"
+    },
+    {
+      "name": "move_note",
+      "description": "Move or rename notes"
+    },
+    {
+      "name": "view_note",
+      "description": "Display notes as formatted artifacts"
+    },
+    {
+      "name": "read_content",
+      "description": "Read raw file content (text, images, binaries)"
+    },
+    {
+      "name": "delete_note",
+      "description": "Delete notes from knowledge base"
+    },
+    {
+      "name": "search_notes",
+      "description": "Full-text search across knowledge base"
+    },
+    {
+      "name": "build_context",
+      "description": "Navigate knowledge graph via memory:// URLs"
+    },
+    {
+      "name": "recent_activity",
+      "description": "Get recently updated information"
+    },
+    {
+      "name": "list_directory",
+      "description": "Browse directory contents with filtering"
+    },
+    {
+      "name": "canvas",
+      "description": "Generate Obsidian canvas files for knowledge visualization"
+    },
+    {
+      "name": "list_memory_projects",
+      "description": "List all projects with status"
+    },
+    {
+      "name": "switch_project",
+      "description": "Switch between project contexts"
+    },
+    {
+      "name": "get_current_project",
+      "description": "Show current project stats"
+    },
+    {
+      "name": "create_memory_project",
+      "description": "Create new projects"
+    },
+    {
+      "name": "delete_project",
+      "description": "Remove projects"
+    },
+    {
+      "name": "set_default_project",
+      "description": "Set default project"
+    },
+    {
+      "name": "sync_status",
+      "description": "Check synchronization status"
+    }
+  ],
+  "prompts": [
+    {
+      "name": "continue_conversation",
+      "description": "Continue previous conversations with relevant historical context",
+      "arguments": [
+        "topic",
+        "timeframe"
+      ],
+      "text": "Continue our conversation about {{topic}} from {{timeframe}} ago. Use the continue_conversation tool to find relevant context and build on our previous discussion."
+    },
+    {
+      "name": "search_notes", 
+      "description": "Search knowledge base with detailed, formatted results",
+      "arguments": [
+        "query",
+        "after_date"
+      ],
+      "text": "Search my knowledge base for information about {{query}}. Use the search_notes tool to find relevant notes and provide a comprehensive summary of what I know about this topic."
+    },
+    {
+      "name": "recent_activity",
+      "description": "View recently changed items with formatted output", 
+      "arguments": [
+        "timeframe"
+      ],
+      "text": "Show me what I've been working on recently in the past {{timeframe}}. Use the recent_activity tool to display my recent notes and changes."
+    }
+  ],
+  "keywords": [
+    "knowledge",
+    "management", 
+    "local-first",
+    "ai",
+    "mcp",
+    "markdown",
+    "notes",
+    "zettelkasten"
+  ],
+  "license": "AGPL-3.0-or-later",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/basicmachines-co/basic-memory"
+  },
+  "compatibility": {
+    "claude_desktop": ">=0.10.0",
+    "platforms": ["darwin", "win32", "linux"],
+    "runtimes": {
+      "python": ">=3.12.0 <4"
+    }
+  }
+}

pyproject.toml

@@ -3,7 +3,7 @@ name = "basic-memory"
 dynamic = ["version"]
 description = "Local-first knowledge management combining Zettelkasten with knowledge graphs"
 readme = "README.md"
-requires-python = ">=3.12"
+requires-python = ">=3.12.1"
 license = { text = "AGPL-3.0-or-later" }
 authors = [
     { name = "Basic Machines", email = "hello@basic-machines.co" }
@@ -14,8 +14,9 @@ dependencies = [
     "typer>=0.9.0",
     "aiosqlite>=0.20.0",
     "greenlet>=3.1.1",
-    "pydantic[email,timezone]>=2.12.0",
-    "mcp>=1.23.1",
+    "pydantic[email,timezone]>=2.10.3",
+    "icecream>=2.1.3",
+    "mcp>=1.2.0",
     "pydantic-settings>=2.6.1",
     "loguru>=0.7.3",
     "pyright>=1.1.390",
@@ -29,26 +30,13 @@ dependencies = [
     "alembic>=1.14.1",
     "pillow>=11.1.0",
     "pybars3>=0.9.7",
-    "fastmcp>=3.0.1,<4",
+    "fastmcp>=2.3.4",
     "pyjwt>=2.10.1",
     "python-dotenv>=1.1.0",
     "pytest-aio>=1.9.0",
-    "aiofiles>=24.1.0", # Optional observability (disabled by default via config)
-    "asyncpg>=0.30.0",
-    "nest-asyncio>=1.6.0", # For Alembic migrations with Postgres
-    "pytest-asyncio>=1.2.0",
-    "psycopg==3.3.1",
-    "mdformat>=0.7.22",
-    "mdformat-gfm>=0.3.7",
-    "mdformat-frontmatter>=2.0.8",
-    "sniffio>=1.3.1",
-    "anyio>=4.10.0",
-    "httpx>=0.28.0",
-    "fastembed>=0.7.4",
-    "sqlite-vec>=0.1.6",
-    "openai>=1.100.2",
 ]
 
+
 [project.urls]
 Homepage = "https://github.com/basicmachines-co/basic-memory"
 Repository = "https://github.com/basicmachines-co/basic-memory"
@@ -64,25 +52,17 @@ build-backend = "hatchling.build"
 
 [tool.pytest.ini_options]
 pythonpath = ["src", "tests"]
-addopts = "--cov=basic_memory --cov-report term-missing"
-testpaths = ["tests", "test-int"]
+addopts = "--cov=basic_memory --cov-report term-missing -ra -q"
+testpaths = ["tests"]
 asyncio_mode = "strict"
 asyncio_default_fixture_loop_scope = "function"
-markers = [
-    "benchmark: Performance benchmark tests (deselect with '-m \"not benchmark\"')",
-    "slow: Slow-running tests (deselect with '-m \"not slow\"')",
-    "postgres: Tests that run against Postgres backend (deselect with '-m \"not postgres\"')",
-    "windows: Windows-specific tests (deselect with '-m \"not windows\"')",
-    "smoke: Fast end-to-end smoke tests for MCP flows",
-    "semantic: Tests requiring semantic dependencies (fastembed, sqlite-vec, openai)",
-]
 
 [tool.ruff]
 line-length = 100
 target-version = "py312"
 
-[dependency-groups]
-dev = [
+[tool.uv]
+dev-dependencies = [
     "gevent>=24.11.1",
     "icecream>=2.1.3",
     "pytest>=8.3.4",
@@ -91,14 +71,6 @@ dev = [
     "pytest-asyncio>=0.24.0",
     "pytest-xdist>=3.0.0",
     "ruff>=0.1.6",
-    "freezegun>=1.5.5",
-    "testcontainers[postgres]>=4.0.0",
-    "psycopg>=3.2.0",
-    "pyright>=1.1.408",
-    "pytest-testmon>=2.2.0",
-    "ty>=0.0.18",
-    "cst-lsp>=0.1.3",
-    "libcst>=1.8.6",
 ]
 
 [tool.hatch.version]
@@ -117,15 +89,12 @@ ignore = ["test/"]
 defineConstant = { DEBUG = true }
 reportMissingImports = "error"
 reportMissingTypeStubs = false
-reportUnusedImport = "none"
 pythonVersion = "3.12"
 
 
 
 [tool.coverage.run]
 concurrency = ["thread", "gevent"]
-parallel = true
-source = ["basic_memory"]
 
 [tool.coverage.report]
 exclude_lines = [
@@ -147,9 +116,11 @@ omit = [
     "*/supabase_auth_provider.py",  # External HTTP calls to Supabase APIs
     "*/watch_service.py",           # File system watching - complex integration testing
     "*/background_sync.py",         # Background processes
-    "*/cli/**",                    # CLI is an interactive wrapper; core logic is covered via API/MCP/service tests
-    "*/db.py",                     # Backend/runtime-dependent (sqlite/postgres/windows tuning); validated via integration tests
-    "*/services/initialization.py", # Startup orchestration + background tasks (watchers); exercised indirectly in entrypoints
-    "*/sync/sync_service.py",      # Heavy filesystem/db integration; covered by integration suite, not enforced in unit coverage
+    "*/cli/main.py",               # CLI entry point
+    "*/mcp/tools/project_management.py",  # Covered by integration tests
+    "*/mcp/tools/sync_status.py",  # Covered by integration tests
     "*/services/migration_service.py", # Complex migration scenarios
 ]
+
+[tool.logfire]
+ignore_no_config = true

server.json

@@ -1,25 +0,0 @@
-{
-  "$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",
-  "name": "io.github.basicmachines-co/basic-memory",
-  "description": "Local-first knowledge management with bi-directional LLM sync via Markdown files.",
-  "repository": {
-    "url": "https://github.com/basicmachines-co/basic-memory.git",
-    "source": "github"
-  },
-  "version": "0.20.1",
-  "packages": [
-    {
-      "registryType": "pypi",
-      "identifier": "basic-memory",
-      "version": "0.20.1",
-      "runtimeHint": "uvx",
-      "runtimeArguments": [
-        {"type": "positional", "value": "basic-memory"},
-        {"type": "positional", "value": "mcp"}
-      ],
-      "transport": {
-        "type": "stdio"
-      }
-    }
-  ]
-}

src/basic_memory/__init__.py

@@ -1,7 +1,7 @@
 """basic-memory - Local-first knowledge management combining Zettelkasten with knowledge graphs"""
 
 # Package version - updated by release automation
-__version__ = "0.20.1"
+__version__ = "0.14.0"
 
 # API version for FastAPI - independent of package version
 __api_version__ = "v0"

src/basic_memory/alembic/env.py

@@ -1,60 +1,29 @@
 """Alembic environment configuration."""
 
-import asyncio
 import os
 from logging.config import fileConfig
 
-# Allow nested event loops (needed for pytest-asyncio and other async contexts)
-# Note: nest_asyncio doesn't work with uvloop or Python 3.14+, so we handle those cases separately
-import sys
-
-if sys.version_info < (3, 14):
-    try:
-        import nest_asyncio
-
-        nest_asyncio.apply()
-    except (ImportError, ValueError):
-        # nest_asyncio not available or can't patch this loop type (e.g., uvloop)
-        pass
-# For Python 3.14+, we rely on the thread-based fallback in run_migrations_online()
-
-from sqlalchemy import engine_from_config, pool
-from sqlalchemy.ext.asyncio import AsyncEngine, create_async_engine
+from sqlalchemy import engine_from_config
+from sqlalchemy import pool
 
 from alembic import context
 
-from basic_memory.config import ConfigManager
-
-# Trigger: only set test env when actually running under pytest
-# Why: alembic/env.py is imported during normal operations (MCP server startup, migrations)
-#      but we only want test behavior during actual test runs
-# Outcome: prevents is_test_env from returning True in production, enabling watch service
-if os.getenv("PYTEST_CURRENT_TEST") is not None:
-    os.environ["BASIC_MEMORY_ENV"] = "test"
+# set config.env to "test" for pytest to prevent logging to file in utils.setup_logging()
+os.environ["BASIC_MEMORY_ENV"] = "test"
 
 # Import after setting environment variable  # noqa: E402
+from basic_memory.config import app_config  # noqa: E402
 from basic_memory.models import Base  # noqa: E402
 
 # this is the Alembic Config object, which provides
 # access to the values within the .ini file in use.
 config = context.config
 
-# Load app config - this will read environment variables (BASIC_MEMORY_DATABASE_BACKEND, etc.)
-# due to Pydantic's env_prefix="BASIC_MEMORY_" setting
-app_config = ConfigManager().config
+# Set the SQLAlchemy URL from our app config
+sqlalchemy_url = f"sqlite:///{app_config.database_path}"
+config.set_main_option("sqlalchemy.url", sqlalchemy_url)
 
-# Set the SQLAlchemy URL based on database backend configuration
-# If the URL is already set in config (e.g., from run_migrations), use that
-# Otherwise, get it from app config
-# Note: alembic.ini has a placeholder URL "driver://user:pass@localhost/dbname" that we need to override
-current_url = config.get_main_option("sqlalchemy.url")
-if not current_url or current_url == "driver://user:pass@localhost/dbname":
-    from basic_memory.db import DatabaseType
-
-    sqlalchemy_url = DatabaseType.get_db_url(
-        app_config.database_path, DatabaseType.FILESYSTEM, app_config
-    )
-    config.set_main_option("sqlalchemy.url", sqlalchemy_url)
+# print(f"Using SQLAlchemy URL: {sqlalchemy_url}")
 
 # Interpret the config file for Python logging.
 if config.config_file_name is not None:
@@ -98,89 +67,28 @@ def run_migrations_offline() -> None:
         context.run_migrations()
 
 
-def do_run_migrations(connection):
-    """Execute migrations with the given connection."""
-    context.configure(
-        connection=connection,
-        target_metadata=target_metadata,
-        include_object=include_object,
-        render_as_batch=True,
-        compare_type=True,
-    )
-    with context.begin_transaction():
-        context.run_migrations()
-
-
-async def run_async_migrations(connectable):
-    """Run migrations asynchronously with AsyncEngine."""
-    async with connectable.connect() as connection:
-        await connection.run_sync(do_run_migrations)
-    await connectable.dispose()
-
-
 def run_migrations_online() -> None:
     """Run migrations in 'online' mode.
 
-    Supports both sync engines (SQLite) and async engines (PostgreSQL with asyncpg).
+    In this scenario we need to create an Engine
+    and associate a connection with the context.
     """
-    # Check if a connection/engine was provided (e.g., from run_migrations)
-    connectable = context.config.attributes.get("connection", None)
+    connectable = engine_from_config(
+        config.get_section(config.config_ini_section, {}),
+        prefix="sqlalchemy.",
+        poolclass=pool.NullPool,
+    )
 
-    if connectable is None:
-        # No connection provided, create engine from config
-        url = context.config.get_main_option("sqlalchemy.url")
+    with connectable.connect() as connection:
+        context.configure(
+            connection=connection,
+            target_metadata=target_metadata,
+            include_object=include_object,
+            render_as_batch=True,
+        )
 
-        # Check if it's an async URL (sqlite+aiosqlite or postgresql+asyncpg)
-        if url and ("+asyncpg" in url or "+aiosqlite" in url):
-            # Create async engine for asyncpg or aiosqlite
-            connectable = create_async_engine(
-                url,
-                poolclass=pool.NullPool,
-                future=True,
-            )
-        else:
-            # Create sync engine for regular sqlite or postgresql
-            connectable = engine_from_config(
-                context.config.get_section(context.config.config_ini_section, {}),
-                prefix="sqlalchemy.",
-                poolclass=pool.NullPool,
-            )
-
-    # Handle async engines (PostgreSQL with asyncpg)
-    if isinstance(connectable, AsyncEngine):
-        # Try to run async migrations
-        # nest_asyncio allows asyncio.run() from within event loops, but doesn't work with uvloop
-        try:
-            asyncio.run(run_async_migrations(connectable))
-        except RuntimeError as e:
-            if "cannot be called from a running event loop" in str(e):
-                # We're in a running event loop (likely uvloop) - need to use a different approach
-                # Create a new thread to run the async migrations
-                import concurrent.futures
-
-                def run_in_thread():
-                    """Run async migrations in a new event loop in a separate thread."""
-                    new_loop = asyncio.new_event_loop()
-                    asyncio.set_event_loop(new_loop)
-                    try:
-                        new_loop.run_until_complete(run_async_migrations(connectable))
-                    finally:
-                        new_loop.close()
-
-                with concurrent.futures.ThreadPoolExecutor() as executor:
-                    future = executor.submit(run_in_thread)
-                    future.result()  # Wait for completion and re-raise any exceptions
-            else:
-                raise
-    else:
-        # Handle sync engines (SQLite) or sync connections
-        if hasattr(connectable, "connect"):
-            # It's an engine, get a connection
-            with connectable.connect() as connection:
-                do_run_migrations(connection)
-        else:
-            # It's already a connection
-            do_run_migrations(connectable)
+        with context.begin_transaction():
+            context.run_migrations()
 
 
 if context.is_offline_mode():

src/basic_memory/alembic/versions/314f1ea54dc4_add_postgres_full_text_search_support_.py

@@ -1,131 +0,0 @@
-"""Add Postgres full-text search support with tsvector and GIN indexes
-
-Revision ID: 314f1ea54dc4
-Revises: e7e1f4367280
-Create Date: 2025-11-15 18:05:01.025405
-
-"""
-
-from typing import Sequence, Union
-
-from alembic import op
-import sqlalchemy as sa
-
-
-# revision identifiers, used by Alembic.
-revision: str = "314f1ea54dc4"
-down_revision: Union[str, None] = "e7e1f4367280"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    """Add PostgreSQL full-text search support.
-
-    This migration:
-    1. Creates search_index table for Postgres (SQLite uses FTS5 virtual table)
-    2. Adds generated tsvector column for full-text search
-    3. Creates GIN index on the tsvector column for fast text queries
-    4. Creates GIN index on metadata JSONB column for fast containment queries
-
-    Note: These changes only apply to Postgres. SQLite continues to use FTS5 virtual tables.
-    """
-    # Check if we're using Postgres
-    connection = op.get_bind()
-    if connection.dialect.name == "postgresql":
-        # Create search_index table for Postgres
-        # For SQLite, this is a FTS5 virtual table created elsewhere
-        from sqlalchemy.dialects.postgresql import JSONB
-
-        op.create_table(
-            "search_index",
-            sa.Column("id", sa.Integer(), nullable=False),  # Entity IDs are integers
-            sa.Column("project_id", sa.Integer(), nullable=False),  # Multi-tenant isolation
-            sa.Column("title", sa.Text(), nullable=True),
-            sa.Column("content_stems", sa.Text(), nullable=True),
-            sa.Column("content_snippet", sa.Text(), nullable=True),
-            sa.Column("permalink", sa.String(), nullable=True),  # Nullable for non-markdown files
-            sa.Column("file_path", sa.String(), nullable=True),
-            sa.Column("type", sa.String(), nullable=True),
-            sa.Column("from_id", sa.Integer(), nullable=True),  # Relation IDs are integers
-            sa.Column("to_id", sa.Integer(), nullable=True),  # Relation IDs are integers
-            sa.Column("relation_type", sa.String(), nullable=True),
-            sa.Column("entity_id", sa.Integer(), nullable=True),  # Entity IDs are integers
-            sa.Column("category", sa.String(), nullable=True),
-            sa.Column("metadata", JSONB(), nullable=True),  # Use JSONB for Postgres
-            sa.Column("created_at", sa.DateTime(timezone=True), nullable=True),
-            sa.Column("updated_at", sa.DateTime(timezone=True), nullable=True),
-            sa.PrimaryKeyConstraint(
-                "id", "type", "project_id"
-            ),  # Composite key: id can repeat across types
-            sa.ForeignKeyConstraint(
-                ["project_id"],
-                ["project.id"],
-                name="fk_search_index_project_id",
-                ondelete="CASCADE",
-            ),
-            if_not_exists=True,
-        )
-
-        # Create index on project_id for efficient multi-tenant queries
-        op.create_index(
-            "ix_search_index_project_id",
-            "search_index",
-            ["project_id"],
-            unique=False,
-        )
-
-        # Create unique partial index on permalink for markdown files
-        # Non-markdown files don't have permalinks, so we use a partial index
-        op.execute("""
-            CREATE UNIQUE INDEX uix_search_index_permalink_project
-            ON search_index (permalink, project_id)
-            WHERE permalink IS NOT NULL
-        """)
-
-        # Add tsvector column as a GENERATED ALWAYS column
-        # This automatically updates when title or content_stems change
-        op.execute("""
-            ALTER TABLE search_index
-            ADD COLUMN textsearchable_index_col tsvector
-            GENERATED ALWAYS AS (
-                to_tsvector('english',
-                    coalesce(title, '') || ' ' ||
-                    coalesce(content_stems, '')
-                )
-            ) STORED
-        """)
-
-        # Create GIN index on tsvector column for fast full-text search
-        op.create_index(
-            "idx_search_index_fts",
-            "search_index",
-            ["textsearchable_index_col"],
-            unique=False,
-            postgresql_using="gin",
-        )
-
-        # Create GIN index on metadata JSONB for fast containment queries
-        # Using jsonb_path_ops for smaller index size and better performance
-        op.execute("""
-            CREATE INDEX idx_search_index_metadata_gin
-            ON search_index
-            USING GIN (metadata jsonb_path_ops)
-        """)
-
-
-def downgrade() -> None:
-    """Remove PostgreSQL full-text search support."""
-    connection = op.get_bind()
-    if connection.dialect.name == "postgresql":
-        # Drop indexes first
-        op.execute("DROP INDEX IF EXISTS idx_search_index_metadata_gin")
-        op.drop_index("idx_search_index_fts", table_name="search_index")
-        op.execute("DROP INDEX IF EXISTS uix_search_index_permalink_project")
-        op.drop_index("ix_search_index_project_id", table_name="search_index")
-
-        # Drop the generated column
-        op.execute("ALTER TABLE search_index DROP COLUMN IF EXISTS textsearchable_index_col")
-
-        # Drop the search_index table
-        op.drop_table("search_index")

src/basic_memory/alembic/versions/5fe1ab1ccebe_add_projects_table.py

@@ -21,12 +21,6 @@ depends_on: Union[str, Sequence[str], None] = None
 
 def upgrade() -> None:
     # ### commands auto generated by Alembic - please adjust! ###
-
-    # SQLite FTS5 virtual table handling is SQLite-specific
-    # For Postgres, search_index is a regular table managed by ORM
-    connection = op.get_bind()
-    is_sqlite = connection.dialect.name == "sqlite"
-
     op.create_table(
         "project",
         sa.Column("id", sa.Integer(), nullable=False),
@@ -61,9 +55,7 @@ def upgrade() -> None:
         batch_op.add_column(sa.Column("project_id", sa.Integer(), nullable=False))
         batch_op.drop_index(
             "uix_entity_permalink",
-            sqlite_where=sa.text("content_type = 'text/markdown' AND permalink IS NOT NULL")
-            if is_sqlite
-            else None,
+            sqlite_where=sa.text("content_type = 'text/markdown' AND permalink IS NOT NULL"),
         )
         batch_op.drop_index("ix_entity_file_path")
         batch_op.create_index(batch_op.f("ix_entity_file_path"), ["file_path"], unique=False)
@@ -75,16 +67,12 @@ def upgrade() -> None:
             "uix_entity_permalink_project",
             ["permalink", "project_id"],
             unique=True,
-            sqlite_where=sa.text("content_type = 'text/markdown' AND permalink IS NOT NULL")
-            if is_sqlite
-            else None,
+            sqlite_where=sa.text("content_type = 'text/markdown' AND permalink IS NOT NULL"),
         )
         batch_op.create_foreign_key("fk_entity_project_id", "project", ["project_id"], ["id"])
 
     # drop the search index table. it will be recreated
-    # Only drop for SQLite - Postgres uses regular table managed by ORM
-    if is_sqlite:
-        op.drop_table("search_index")
+    op.drop_table("search_index")
 
     # ### end Alembic commands ###
 

src/basic_memory/alembic/versions/647e7a75e2cd_project_constraint_fix.py

@@ -25,51 +25,43 @@ def upgrade() -> None:
     The UNIQUE constraint prevents multiple projects from having is_default=FALSE,
     which breaks project creation when the service sets is_default=False.
 
-    SQLite: Recreate the table without the constraint (no ALTER TABLE support)
-    Postgres: Use ALTER TABLE to drop the constraint directly
+    Since SQLite doesn't support dropping specific constraints easily, we'll
+    recreate the table without the problematic constraint.
     """
-    connection = op.get_bind()
-    is_sqlite = connection.dialect.name == "sqlite"
+    # For SQLite, we need to recreate the table without the UNIQUE constraint
+    # Create a new table without the UNIQUE constraint on is_default
+    op.create_table(
+        "project_new",
+        sa.Column("id", sa.Integer(), nullable=False),
+        sa.Column("name", sa.String(), nullable=False),
+        sa.Column("description", sa.Text(), nullable=True),
+        sa.Column("permalink", sa.String(), nullable=False),
+        sa.Column("path", sa.String(), nullable=False),
+        sa.Column("is_active", sa.Boolean(), nullable=False),
+        sa.Column("is_default", sa.Boolean(), nullable=True),  # No UNIQUE constraint!
+        sa.Column("created_at", sa.DateTime(), nullable=False),
+        sa.Column("updated_at", sa.DateTime(), nullable=False),
+        sa.PrimaryKeyConstraint("id"),
+        sa.UniqueConstraint("name"),
+        sa.UniqueConstraint("permalink"),
+    )
 
-    if is_sqlite:
-        # For SQLite, we need to recreate the table without the UNIQUE constraint
-        # Create a new table without the UNIQUE constraint on is_default
-        op.create_table(
-            "project_new",
-            sa.Column("id", sa.Integer(), nullable=False),
-            sa.Column("name", sa.String(), nullable=False),
-            sa.Column("description", sa.Text(), nullable=True),
-            sa.Column("permalink", sa.String(), nullable=False),
-            sa.Column("path", sa.String(), nullable=False),
-            sa.Column("is_active", sa.Boolean(), nullable=False),
-            sa.Column("is_default", sa.Boolean(), nullable=True),  # No UNIQUE constraint!
-            sa.Column("created_at", sa.DateTime(), nullable=False),
-            sa.Column("updated_at", sa.DateTime(), nullable=False),
-            sa.PrimaryKeyConstraint("id"),
-            sa.UniqueConstraint("name"),
-            sa.UniqueConstraint("permalink"),
-        )
+    # Copy data from old table to new table
+    op.execute("INSERT INTO project_new SELECT * FROM project")
 
-        # Copy data from old table to new table
-        op.execute("INSERT INTO project_new SELECT * FROM project")
+    # Drop the old table
+    op.drop_table("project")
 
-        # Drop the old table
-        op.drop_table("project")
+    # Rename the new table
+    op.rename_table("project_new", "project")
 
-        # Rename the new table
-        op.rename_table("project_new", "project")
-
-        # Recreate the indexes
-        with op.batch_alter_table("project", schema=None) as batch_op:
-            batch_op.create_index("ix_project_created_at", ["created_at"], unique=False)
-            batch_op.create_index("ix_project_name", ["name"], unique=True)
-            batch_op.create_index("ix_project_path", ["path"], unique=False)
-            batch_op.create_index("ix_project_permalink", ["permalink"], unique=True)
-            batch_op.create_index("ix_project_updated_at", ["updated_at"], unique=False)
-    else:
-        # For Postgres, we can simply drop the constraint
-        with op.batch_alter_table("project", schema=None) as batch_op:
-            batch_op.drop_constraint("project_is_default_key", type_="unique")
+    # Recreate the indexes
+    with op.batch_alter_table("project", schema=None) as batch_op:
+        batch_op.create_index("ix_project_created_at", ["created_at"], unique=False)
+        batch_op.create_index("ix_project_name", ["name"], unique=True)
+        batch_op.create_index("ix_project_path", ["path"], unique=False)
+        batch_op.create_index("ix_project_permalink", ["permalink"], unique=True)
+        batch_op.create_index("ix_project_updated_at", ["updated_at"], unique=False)
 
 
 def downgrade() -> None:

src/basic_memory/alembic/versions/6830751f5fb6_merge_multiple_heads.py

@@ -1,24 +0,0 @@
-"""Merge multiple heads
-
-Revision ID: 6830751f5fb6
-Revises: a2b3c4d5e6f7, g9a0b3c4d5e6
-Create Date: 2025-12-29 12:46:46.476268
-
-"""
-
-from typing import Sequence, Union
-
-
-# revision identifiers, used by Alembic.
-revision: str = "6830751f5fb6"
-down_revision: Union[str, Sequence[str], None] = ("a2b3c4d5e6f7", "g9a0b3c4d5e6")
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    pass
-
-
-def downgrade() -> None:
-    pass

src/basic_memory/alembic/versions/9d9c1cb7d8f5_add_mtime_and_size_columns_to_entity_.py

@@ -1,49 +0,0 @@
-"""Add mtime and size columns to Entity for sync optimization
-
-Revision ID: 9d9c1cb7d8f5
-Revises: a1b2c3d4e5f6
-Create Date: 2025-10-20 05:07:55.173849
-
-"""
-
-from typing import Sequence, Union
-
-from alembic import op
-import sqlalchemy as sa
-
-
-# revision identifiers, used by Alembic.
-revision: str = "9d9c1cb7d8f5"
-down_revision: Union[str, None] = "a1b2c3d4e5f6"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    # ### commands auto generated by Alembic - please adjust! ###
-    with op.batch_alter_table("entity", schema=None) as batch_op:
-        batch_op.add_column(sa.Column("mtime", sa.Float(), nullable=True))
-        batch_op.add_column(sa.Column("size", sa.Integer(), nullable=True))
-        batch_op.drop_constraint(batch_op.f("fk_entity_project_id"), type_="foreignkey")
-        batch_op.create_foreign_key(
-            batch_op.f("fk_entity_project_id"), "project", ["project_id"], ["id"]
-        )
-
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    # ### commands auto generated by Alembic - please adjust! ###
-    with op.batch_alter_table("entity", schema=None) as batch_op:
-        batch_op.drop_constraint(batch_op.f("fk_entity_project_id"), type_="foreignkey")
-        batch_op.create_foreign_key(
-            batch_op.f("fk_entity_project_id"),
-            "project",
-            ["project_id"],
-            ["id"],
-            ondelete="CASCADE",
-        )
-        batch_op.drop_column("size")
-        batch_op.drop_column("mtime")
-
-    # ### end Alembic commands ###

src/basic_memory/alembic/versions/a1b2c3d4e5f6_fix_project_foreign_keys.py

@@ -1,49 +0,0 @@
-"""fix project foreign keys
-
-Revision ID: a1b2c3d4e5f6
-Revises: 647e7a75e2cd
-Create Date: 2025-08-19 22:06:00.000000
-
-"""
-
-from typing import Sequence, Union
-
-from alembic import op
-
-
-# revision identifiers, used by Alembic.
-revision: str = "a1b2c3d4e5f6"
-down_revision: Union[str, None] = "647e7a75e2cd"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    """Re-establish foreign key constraints that were lost during project table recreation.
-
-    The migration 647e7a75e2cd recreated the project table but did not re-establish
-    the foreign key constraint from entity.project_id to project.id, causing
-    foreign key constraint failures when trying to delete projects with related entities.
-    """
-    # SQLite doesn't allow adding foreign key constraints to existing tables easily
-    # We need to be careful and handle the case where the constraint might already exist
-
-    with op.batch_alter_table("entity", schema=None) as batch_op:
-        # Try to drop existing foreign key constraint (may not exist)
-        try:
-            batch_op.drop_constraint("fk_entity_project_id", type_="foreignkey")
-        except Exception:
-            # Constraint may not exist, which is fine - we'll create it next
-            pass
-
-        # Add the foreign key constraint with CASCADE DELETE
-        # This ensures that when a project is deleted, all related entities are also deleted
-        batch_op.create_foreign_key(
-            "fk_entity_project_id", "project", ["project_id"], ["id"], ondelete="CASCADE"
-        )
-
-
-def downgrade() -> None:
-    """Remove the foreign key constraint."""
-    with op.batch_alter_table("entity", schema=None) as batch_op:
-        batch_op.drop_constraint("fk_entity_project_id", type_="foreignkey")

src/basic_memory/alembic/versions/a2b3c4d5e6f7_add_search_index_entity_cascade.py

@@ -1,56 +0,0 @@
-"""Add cascade delete FK from search_index to entity
-
-Revision ID: a2b3c4d5e6f7
-Revises: f8a9b2c3d4e5
-Create Date: 2025-12-02 07:00:00.000000
-
-"""
-
-from typing import Sequence, Union
-
-from alembic import op
-
-
-# revision identifiers, used by Alembic.
-revision: str = "a2b3c4d5e6f7"
-down_revision: Union[str, None] = "f8a9b2c3d4e5"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    """Add FK with CASCADE delete from search_index.entity_id to entity.id.
-
-    This migration is Postgres-only because:
-    - SQLite uses FTS5 virtual tables which don't support foreign keys
-    - The FK enables automatic cleanup of search_index entries when entities are deleted
-    """
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    if dialect == "postgresql":
-        # First, clean up any orphaned search_index entries where entity no longer exists
-        op.execute("""
-            DELETE FROM search_index
-            WHERE entity_id IS NOT NULL
-            AND entity_id NOT IN (SELECT id FROM entity)
-        """)
-
-        # Add FK with CASCADE - nullable FK allows search_index entries without entity_id
-        op.create_foreign_key(
-            "fk_search_index_entity_id",
-            "search_index",
-            "entity",
-            ["entity_id"],
-            ["id"],
-            ondelete="CASCADE",
-        )
-
-
-def downgrade() -> None:
-    """Remove the FK constraint."""
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    if dialect == "postgresql":
-        op.drop_constraint("fk_search_index_entity_id", "search_index", type_="foreignkey")

src/basic_memory/alembic/versions/cc7172b46608_update_search_index_schema.py

@@ -21,12 +21,6 @@ depends_on: Union[str, Sequence[str], None] = None
 def upgrade() -> None:
     """Upgrade database schema to use new search index with content_stems and content_snippet."""
 
-    # This migration is SQLite-specific (FTS5 virtual tables)
-    # For Postgres, the search_index table is created via ORM models
-    connection = op.get_bind()
-    if connection.dialect.name != "sqlite":
-        return
-
     # First, drop the existing search_index table
     op.execute("DROP TABLE IF EXISTS search_index")
 
@@ -65,13 +59,6 @@ def upgrade() -> None:
 
 def downgrade() -> None:
     """Downgrade database schema to use old search index."""
-
-    # This migration is SQLite-specific (FTS5 virtual tables)
-    # For Postgres, the search_index table is managed via ORM models
-    connection = op.get_bind()
-    if connection.dialect.name != "sqlite":
-        return
-
     # Drop the updated search_index table
     op.execute("DROP TABLE IF EXISTS search_index")
 

src/basic_memory/alembic/versions/d7e8f9a0b1c2_add_structured_metadata_indexes.py

@@ -1,154 +0,0 @@
-"""Add structured metadata indexes for entity frontmatter
-
-Revision ID: d7e8f9a0b1c2
-Revises: g9a0b3c4d5e6
-Create Date: 2026-01-31 12:00:00.000000
-
-"""
-
-from typing import Sequence, Union
-
-import sqlalchemy as sa
-from alembic import op
-from sqlalchemy import text
-
-
-def column_exists(connection, table: str, column: str) -> bool:
-    """Check if a column exists in a table (idempotent migration support)."""
-    if connection.dialect.name == "postgresql":
-        result = connection.execute(
-            text(
-                "SELECT 1 FROM information_schema.columns "
-                "WHERE table_name = :table AND column_name = :column"
-            ),
-            {"table": table, "column": column},
-        )
-        return result.fetchone() is not None
-    # SQLite
-    result = connection.execute(text(f"PRAGMA table_info({table})"))
-    columns = [row[1] for row in result]
-    return column in columns
-
-
-def index_exists(connection, index_name: str) -> bool:
-    """Check if an index exists (idempotent migration support)."""
-    if connection.dialect.name == "postgresql":
-        result = connection.execute(
-            text("SELECT 1 FROM pg_indexes WHERE indexname = :index_name"),
-            {"index_name": index_name},
-        )
-        return result.fetchone() is not None
-    # SQLite
-    result = connection.execute(
-        text("SELECT 1 FROM sqlite_master WHERE type='index' AND name = :index_name"),
-        {"index_name": index_name},
-    )
-    return result.fetchone() is not None
-
-
-# revision identifiers, used by Alembic.
-revision: str = "d7e8f9a0b1c2"
-down_revision: Union[str, None] = "6830751f5fb6"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    """Add JSONB/GiN indexes for Postgres and generated columns for SQLite."""
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    if dialect == "postgresql":
-        # Ensure JSONB for efficient indexing
-        result = connection.execute(
-            text(
-                "SELECT data_type FROM information_schema.columns "
-                "WHERE table_name = 'entity' AND column_name = 'entity_metadata'"
-            )
-        ).fetchone()
-        if result and result[0] != "jsonb":
-            op.execute(
-                "ALTER TABLE entity ALTER COLUMN entity_metadata "
-                "TYPE jsonb USING entity_metadata::jsonb"
-            )
-
-        # General JSONB GIN index
-        op.execute(
-            "CREATE INDEX IF NOT EXISTS idx_entity_metadata_gin "
-            "ON entity USING GIN (entity_metadata jsonb_path_ops)"
-        )
-
-        # Common field indexes
-        op.execute(
-            "CREATE INDEX IF NOT EXISTS idx_entity_tags_json "
-            "ON entity USING GIN ((entity_metadata -> 'tags'))"
-        )
-        op.execute(
-            "CREATE INDEX IF NOT EXISTS idx_entity_frontmatter_type "
-            "ON entity ((entity_metadata ->> 'type'))"
-        )
-        op.execute(
-            "CREATE INDEX IF NOT EXISTS idx_entity_frontmatter_status "
-            "ON entity ((entity_metadata ->> 'status'))"
-        )
-        return
-
-    # SQLite: add generated columns for common frontmatter fields
-    # Constraint: SQLite ALTER TABLE ADD COLUMN only supports VIRTUAL generated columns,
-    # not STORED. json_extract is deterministic so VIRTUAL columns can still be indexed.
-    if not column_exists(connection, "entity", "tags_json"):
-        op.add_column(
-            "entity",
-            sa.Column(
-                "tags_json",
-                sa.Text(),
-                sa.Computed("json_extract(entity_metadata, '$.tags')", persisted=False),
-            ),
-        )
-    if not column_exists(connection, "entity", "frontmatter_status"):
-        op.add_column(
-            "entity",
-            sa.Column(
-                "frontmatter_status",
-                sa.Text(),
-                sa.Computed("json_extract(entity_metadata, '$.status')", persisted=False),
-            ),
-        )
-    if not column_exists(connection, "entity", "frontmatter_type"):
-        op.add_column(
-            "entity",
-            sa.Column(
-                "frontmatter_type",
-                sa.Text(),
-                sa.Computed("json_extract(entity_metadata, '$.type')", persisted=False),
-            ),
-        )
-
-    # Index generated columns
-    if not index_exists(connection, "idx_entity_tags_json"):
-        op.create_index("idx_entity_tags_json", "entity", ["tags_json"])
-    if not index_exists(connection, "idx_entity_frontmatter_status"):
-        op.create_index("idx_entity_frontmatter_status", "entity", ["frontmatter_status"])
-    if not index_exists(connection, "idx_entity_frontmatter_type"):
-        op.create_index("idx_entity_frontmatter_type", "entity", ["frontmatter_type"])
-
-
-def downgrade() -> None:
-    """Best-effort downgrade (drop indexes, revert JSONB on Postgres)."""
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    if dialect == "postgresql":
-        op.execute("DROP INDEX IF EXISTS idx_entity_frontmatter_status")
-        op.execute("DROP INDEX IF EXISTS idx_entity_frontmatter_type")
-        op.execute("DROP INDEX IF EXISTS idx_entity_tags_json")
-        op.execute("DROP INDEX IF EXISTS idx_entity_metadata_gin")
-        op.execute(
-            "ALTER TABLE entity ALTER COLUMN entity_metadata TYPE json USING entity_metadata::json"
-        )
-        return
-
-    # SQLite: drop indexes (dropping generated columns requires table rebuild)
-    op.execute("DROP INDEX IF EXISTS idx_entity_frontmatter_status")
-    op.execute("DROP INDEX IF EXISTS idx_entity_frontmatter_type")
-    op.execute("DROP INDEX IF EXISTS idx_entity_tags_json")

src/basic_memory/alembic/versions/e7e1f4367280_add_scan_watermark_tracking_to_project.py

@@ -1,37 +0,0 @@
-"""Add scan watermark tracking to Project
-
-Revision ID: e7e1f4367280
-Revises: 9d9c1cb7d8f5
-Create Date: 2025-10-20 16:42:46.625075
-
-"""
-
-from typing import Sequence, Union
-
-from alembic import op
-import sqlalchemy as sa
-
-
-# revision identifiers, used by Alembic.
-revision: str = "e7e1f4367280"
-down_revision: Union[str, None] = "9d9c1cb7d8f5"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    # ### commands auto generated by Alembic - please adjust! ###
-    with op.batch_alter_table("project", schema=None) as batch_op:
-        batch_op.add_column(sa.Column("last_scan_timestamp", sa.Float(), nullable=True))
-        batch_op.add_column(sa.Column("last_file_count", sa.Integer(), nullable=True))
-
-    # ### end Alembic commands ###
-
-
-def downgrade() -> None:
-    # ### commands auto generated by Alembic - please adjust! ###
-    with op.batch_alter_table("project", schema=None) as batch_op:
-        batch_op.drop_column("last_file_count")
-        batch_op.drop_column("last_scan_timestamp")
-
-    # ### end Alembic commands ###

src/basic_memory/alembic/versions/f8a9b2c3d4e5_add_pg_trgm_for_fuzzy_link_resolution.py

@@ -1,239 +0,0 @@
-"""Add project_id to relation/observation and pg_trgm for fuzzy link resolution
-
-Revision ID: f8a9b2c3d4e5
-Revises: 314f1ea54dc4
-Create Date: 2025-12-01 12:00:00.000000
-
-"""
-
-from typing import Sequence, Union
-
-import sqlalchemy as sa
-from alembic import op
-from sqlalchemy import text
-
-
-def column_exists(connection, table: str, column: str) -> bool:
-    """Check if a column exists in a table (idempotent migration support)."""
-    if connection.dialect.name == "postgresql":
-        result = connection.execute(
-            text(
-                "SELECT 1 FROM information_schema.columns "
-                "WHERE table_name = :table AND column_name = :column"
-            ),
-            {"table": table, "column": column},
-        )
-        return result.fetchone() is not None
-    else:
-        # SQLite
-        result = connection.execute(text(f"PRAGMA table_info({table})"))
-        columns = [row[1] for row in result]
-        return column in columns
-
-
-def index_exists(connection, index_name: str) -> bool:
-    """Check if an index exists (idempotent migration support)."""
-    if connection.dialect.name == "postgresql":
-        result = connection.execute(
-            text("SELECT 1 FROM pg_indexes WHERE indexname = :index_name"),
-            {"index_name": index_name},
-        )
-        return result.fetchone() is not None
-    else:
-        # SQLite
-        result = connection.execute(
-            text("SELECT 1 FROM sqlite_master WHERE type='index' AND name = :index_name"),
-            {"index_name": index_name},
-        )
-        return result.fetchone() is not None
-
-
-# revision identifiers, used by Alembic.
-revision: str = "f8a9b2c3d4e5"
-down_revision: Union[str, None] = "314f1ea54dc4"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    """Add project_id to relation and observation tables, plus pg_trgm indexes.
-
-    This migration:
-    1. Adds project_id column to relation and observation tables (denormalization)
-    2. Backfills project_id from the associated entity
-    3. Enables pg_trgm extension for trigram-based fuzzy matching (Postgres only)
-    4. Creates GIN indexes on entity title and permalink for fast similarity searches
-    5. Creates partial index on unresolved relations for efficient bulk resolution
-    """
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    # -------------------------------------------------------------------------
-    # Add project_id to relation table
-    # -------------------------------------------------------------------------
-
-    # Step 1: Add project_id column as nullable first (idempotent)
-    if not column_exists(connection, "relation", "project_id"):
-        op.add_column("relation", sa.Column("project_id", sa.Integer(), nullable=True))
-
-        # Step 2: Backfill project_id from entity.project_id via from_id
-        if dialect == "postgresql":
-            op.execute("""
-                UPDATE relation
-                SET project_id = entity.project_id
-                FROM entity
-                WHERE relation.from_id = entity.id
-            """)
-        else:
-            # SQLite syntax
-            op.execute("""
-                UPDATE relation
-                SET project_id = (
-                    SELECT entity.project_id
-                    FROM entity
-                    WHERE entity.id = relation.from_id
-                )
-            """)
-
-        # Step 3: Make project_id NOT NULL and add foreign key
-        if dialect == "postgresql":
-            op.alter_column("relation", "project_id", nullable=False)
-            op.create_foreign_key(
-                "fk_relation_project_id",
-                "relation",
-                "project",
-                ["project_id"],
-                ["id"],
-            )
-        else:
-            # SQLite requires batch operations for ALTER COLUMN
-            with op.batch_alter_table("relation") as batch_op:
-                batch_op.alter_column("project_id", nullable=False)
-                batch_op.create_foreign_key(
-                    "fk_relation_project_id",
-                    "project",
-                    ["project_id"],
-                    ["id"],
-                )
-
-    # Step 4: Create index on relation.project_id (idempotent)
-    if not index_exists(connection, "ix_relation_project_id"):
-        op.create_index("ix_relation_project_id", "relation", ["project_id"])
-
-    # -------------------------------------------------------------------------
-    # Add project_id to observation table
-    # -------------------------------------------------------------------------
-
-    # Step 1: Add project_id column as nullable first (idempotent)
-    if not column_exists(connection, "observation", "project_id"):
-        op.add_column("observation", sa.Column("project_id", sa.Integer(), nullable=True))
-
-        # Step 2: Backfill project_id from entity.project_id via entity_id
-        if dialect == "postgresql":
-            op.execute("""
-                UPDATE observation
-                SET project_id = entity.project_id
-                FROM entity
-                WHERE observation.entity_id = entity.id
-            """)
-        else:
-            # SQLite syntax
-            op.execute("""
-                UPDATE observation
-                SET project_id = (
-                    SELECT entity.project_id
-                    FROM entity
-                    WHERE entity.id = observation.entity_id
-                )
-            """)
-
-        # Step 3: Make project_id NOT NULL and add foreign key
-        if dialect == "postgresql":
-            op.alter_column("observation", "project_id", nullable=False)
-            op.create_foreign_key(
-                "fk_observation_project_id",
-                "observation",
-                "project",
-                ["project_id"],
-                ["id"],
-            )
-        else:
-            # SQLite requires batch operations for ALTER COLUMN
-            with op.batch_alter_table("observation") as batch_op:
-                batch_op.alter_column("project_id", nullable=False)
-                batch_op.create_foreign_key(
-                    "fk_observation_project_id",
-                    "project",
-                    ["project_id"],
-                    ["id"],
-                )
-
-    # Step 4: Create index on observation.project_id (idempotent)
-    if not index_exists(connection, "ix_observation_project_id"):
-        op.create_index("ix_observation_project_id", "observation", ["project_id"])
-
-    # Postgres-specific: pg_trgm and GIN indexes
-    if dialect == "postgresql":
-        # Enable pg_trgm extension for fuzzy string matching
-        op.execute("CREATE EXTENSION IF NOT EXISTS pg_trgm")
-
-        # Create trigram indexes on entity table for fuzzy matching
-        # GIN indexes with gin_trgm_ops support similarity searches
-        op.execute("""
-            CREATE INDEX IF NOT EXISTS idx_entity_title_trgm
-            ON entity USING gin (title gin_trgm_ops)
-        """)
-
-        op.execute("""
-            CREATE INDEX IF NOT EXISTS idx_entity_permalink_trgm
-            ON entity USING gin (permalink gin_trgm_ops)
-        """)
-
-        # Create partial index on unresolved relations for efficient bulk resolution
-        # This makes "WHERE to_id IS NULL AND project_id = X" queries very fast
-        op.execute("""
-            CREATE INDEX IF NOT EXISTS idx_relation_unresolved
-            ON relation (project_id, to_name)
-            WHERE to_id IS NULL
-        """)
-
-        # Create index on relation.to_name for join performance in bulk resolution
-        op.execute("""
-            CREATE INDEX IF NOT EXISTS idx_relation_to_name
-            ON relation (to_name)
-        """)
-
-
-def downgrade() -> None:
-    """Remove project_id from relation/observation and pg_trgm indexes."""
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    if dialect == "postgresql":
-        # Drop Postgres-specific indexes
-        op.execute("DROP INDEX IF EXISTS idx_relation_to_name")
-        op.execute("DROP INDEX IF EXISTS idx_relation_unresolved")
-        op.execute("DROP INDEX IF EXISTS idx_entity_permalink_trgm")
-        op.execute("DROP INDEX IF EXISTS idx_entity_title_trgm")
-        # Note: We don't drop the pg_trgm extension as other code may depend on it
-
-        # Drop project_id from observation
-        op.drop_index("ix_observation_project_id", table_name="observation")
-        op.drop_constraint("fk_observation_project_id", "observation", type_="foreignkey")
-        op.drop_column("observation", "project_id")
-
-        # Drop project_id from relation
-        op.drop_index("ix_relation_project_id", table_name="relation")
-        op.drop_constraint("fk_relation_project_id", "relation", type_="foreignkey")
-        op.drop_column("relation", "project_id")
-    else:
-        # SQLite requires batch operations
-        op.drop_index("ix_observation_project_id", table_name="observation")
-        with op.batch_alter_table("observation") as batch_op:
-            batch_op.drop_constraint("fk_observation_project_id", type_="foreignkey")
-            batch_op.drop_column("project_id")
-
-        op.drop_index("ix_relation_project_id", table_name="relation")
-        with op.batch_alter_table("relation") as batch_op:
-            batch_op.drop_constraint("fk_relation_project_id", type_="foreignkey")
-            batch_op.drop_column("project_id")

src/basic_memory/alembic/versions/g9a0b3c4d5e6_add_external_id_to_project_and_entity.py

@@ -1,173 +0,0 @@
-"""Add external_id UUID column to project and entity tables
-
-Revision ID: g9a0b3c4d5e6
-Revises: f8a9b2c3d4e5
-Create Date: 2025-12-29 10:00:00.000000
-
-"""
-
-import uuid
-from typing import Sequence, Union
-
-import sqlalchemy as sa
-from alembic import op
-from sqlalchemy import text
-
-
-def column_exists(connection, table: str, column: str) -> bool:
-    """Check if a column exists in a table (idempotent migration support)."""
-    if connection.dialect.name == "postgresql":
-        result = connection.execute(
-            text(
-                "SELECT 1 FROM information_schema.columns "
-                "WHERE table_name = :table AND column_name = :column"
-            ),
-            {"table": table, "column": column},
-        )
-        return result.fetchone() is not None
-    else:
-        # SQLite
-        result = connection.execute(text(f"PRAGMA table_info({table})"))
-        columns = [row[1] for row in result]
-        return column in columns
-
-
-def index_exists(connection, index_name: str) -> bool:
-    """Check if an index exists (idempotent migration support)."""
-    if connection.dialect.name == "postgresql":
-        result = connection.execute(
-            text("SELECT 1 FROM pg_indexes WHERE indexname = :index_name"),
-            {"index_name": index_name},
-        )
-        return result.fetchone() is not None
-    else:
-        # SQLite
-        result = connection.execute(
-            text("SELECT 1 FROM sqlite_master WHERE type='index' AND name = :index_name"),
-            {"index_name": index_name},
-        )
-        return result.fetchone() is not None
-
-
-# revision identifiers, used by Alembic.
-revision: str = "g9a0b3c4d5e6"
-down_revision: Union[str, None] = "f8a9b2c3d4e5"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    """Add external_id UUID column to project and entity tables.
-
-    This migration:
-    1. Adds external_id column to project table
-    2. Adds external_id column to entity table
-    3. Generates UUIDs for existing rows
-    4. Creates unique indexes on both columns
-    """
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    # -------------------------------------------------------------------------
-    # Add external_id to project table
-    # -------------------------------------------------------------------------
-
-    if not column_exists(connection, "project", "external_id"):
-        # Step 1: Add external_id column as nullable first
-        op.add_column("project", sa.Column("external_id", sa.String(), nullable=True))
-
-        # Step 2: Generate UUIDs for existing rows
-        if dialect == "postgresql":
-            # Postgres has gen_random_uuid() function
-            op.execute("""
-                UPDATE project
-                SET external_id = gen_random_uuid()::text
-                WHERE external_id IS NULL
-            """)
-        else:
-            # SQLite: need to generate UUIDs in Python
-            result = connection.execute(text("SELECT id FROM project WHERE external_id IS NULL"))
-            for row in result:
-                new_uuid = str(uuid.uuid4())
-                connection.execute(
-                    text("UPDATE project SET external_id = :uuid WHERE id = :id"),
-                    {"uuid": new_uuid, "id": row[0]},
-                )
-
-        # Step 3: Make external_id NOT NULL
-        if dialect == "postgresql":
-            op.alter_column("project", "external_id", nullable=False)
-        else:
-            # SQLite requires batch operations for ALTER COLUMN
-            with op.batch_alter_table("project") as batch_op:
-                batch_op.alter_column("external_id", nullable=False)
-
-    # Step 4: Create unique index on project.external_id (idempotent)
-    if not index_exists(connection, "ix_project_external_id"):
-        op.create_index("ix_project_external_id", "project", ["external_id"], unique=True)
-
-    # -------------------------------------------------------------------------
-    # Add external_id to entity table
-    # -------------------------------------------------------------------------
-
-    if not column_exists(connection, "entity", "external_id"):
-        # Step 1: Add external_id column as nullable first
-        op.add_column("entity", sa.Column("external_id", sa.String(), nullable=True))
-
-        # Step 2: Generate UUIDs for existing rows
-        if dialect == "postgresql":
-            # Postgres has gen_random_uuid() function
-            op.execute("""
-                UPDATE entity
-                SET external_id = gen_random_uuid()::text
-                WHERE external_id IS NULL
-            """)
-        else:
-            # SQLite: need to generate UUIDs in Python
-            result = connection.execute(text("SELECT id FROM entity WHERE external_id IS NULL"))
-            for row in result:
-                new_uuid = str(uuid.uuid4())
-                connection.execute(
-                    text("UPDATE entity SET external_id = :uuid WHERE id = :id"),
-                    {"uuid": new_uuid, "id": row[0]},
-                )
-
-        # Step 3: Make external_id NOT NULL
-        if dialect == "postgresql":
-            op.alter_column("entity", "external_id", nullable=False)
-        else:
-            # SQLite requires batch operations for ALTER COLUMN
-            with op.batch_alter_table("entity") as batch_op:
-                batch_op.alter_column("external_id", nullable=False)
-
-    # Step 4: Create unique index on entity.external_id (idempotent)
-    if not index_exists(connection, "ix_entity_external_id"):
-        op.create_index("ix_entity_external_id", "entity", ["external_id"], unique=True)
-
-
-def downgrade() -> None:
-    """Remove external_id columns from project and entity tables."""
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    # Drop from entity table
-    if index_exists(connection, "ix_entity_external_id"):
-        op.drop_index("ix_entity_external_id", table_name="entity")
-
-    if column_exists(connection, "entity", "external_id"):
-        if dialect == "postgresql":
-            op.drop_column("entity", "external_id")
-        else:
-            with op.batch_alter_table("entity") as batch_op:
-                batch_op.drop_column("external_id")
-
-    # Drop from project table
-    if index_exists(connection, "ix_project_external_id"):
-        op.drop_index("ix_project_external_id", table_name="project")
-
-    if column_exists(connection, "project", "external_id"):
-        if dialect == "postgresql":
-            op.drop_column("project", "external_id")
-        else:
-            with op.batch_alter_table("project") as batch_op:
-                batch_op.drop_column("external_id")

src/basic_memory/alembic/versions/h1b2c3d4e5f6_add_postgres_vector_search_tables.py

@@ -1,68 +0,0 @@
-"""Add Postgres semantic vector search tables (pgvector-aware, optional)
-
-Revision ID: h1b2c3d4e5f6
-Revises: d7e8f9a0b1c2
-Create Date: 2026-02-07 00:00:00.000000
-
-"""
-
-from typing import Sequence, Union
-
-from alembic import op
-
-# revision identifiers, used by Alembic.
-revision: str = "h1b2c3d4e5f6"
-down_revision: Union[str, None] = "d7e8f9a0b1c2"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    """Create Postgres vector chunk metadata table.
-
-    Trigger: database backend is PostgreSQL.
-    Why: search_vector_chunks stores text metadata with no vector-dimension
-    dependency, so it's safe in a migration.  search_vector_embeddings (which
-    requires pgvector and a provider-specific dimension) is created at runtime
-    by PostgresSearchRepository._ensure_vector_tables(), mirroring the SQLite
-    pattern where vector tables are created dynamically.
-    Outcome: creates the dimension-independent chunks table.  The embeddings
-    table + HNSW index are deferred to runtime.
-    """
-    connection = op.get_bind()
-    if connection.dialect.name != "postgresql":
-        return
-
-    op.execute(
-        """
-        CREATE TABLE IF NOT EXISTS search_vector_chunks (
-            id BIGSERIAL PRIMARY KEY,
-            entity_id INTEGER NOT NULL,
-            project_id INTEGER NOT NULL,
-            chunk_key TEXT NOT NULL,
-            chunk_text TEXT NOT NULL,
-            source_hash TEXT NOT NULL,
-            updated_at TIMESTAMPTZ NOT NULL DEFAULT NOW(),
-            UNIQUE (project_id, entity_id, chunk_key)
-        )
-        """
-    )
-    op.execute(
-        """
-        CREATE INDEX IF NOT EXISTS idx_search_vector_chunks_project_entity
-        ON search_vector_chunks (project_id, entity_id)
-        """
-    )
-
-
-def downgrade() -> None:
-    """Remove Postgres vector chunk/embedding tables.
-
-    Does not drop pgvector extension because other schema objects may depend on it.
-    """
-    connection = op.get_bind()
-    if connection.dialect.name != "postgresql":
-        return
-
-    op.execute("DROP TABLE IF EXISTS search_vector_embeddings")
-    op.execute("DROP TABLE IF EXISTS search_vector_chunks")

src/basic_memory/alembic/versions/i2c3d4e5f6g7_auto_backfill_semantic_embeddings.py

@@ -1,29 +0,0 @@
-"""Trigger automatic semantic embedding backfill during migration.
-
-Revision ID: i2c3d4e5f6g7
-Revises: h1b2c3d4e5f6
-Create Date: 2026-02-19 00:00:00.000000
-
-"""
-
-from typing import Sequence, Union
-
-# revision identifiers, used by Alembic.
-revision: str = "i2c3d4e5f6g7"
-down_revision: Union[str, None] = "h1b2c3d4e5f6"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def upgrade() -> None:
-    """No schema change.
-
-    Trigger: this revision is newly applied.
-    Why: db.run_migrations() detects this revision transition and runs the existing
-    sync_entity_vectors() pipeline to backfill semantic embeddings automatically.
-    Outcome: users no longer need to run `bm reindex --embeddings` after upgrading.
-    """
-
-
-def downgrade() -> None:
-    """No-op downgrade."""

src/basic_memory/alembic/versions/j3d4e5f6g7h8_rename_entity_type_to_note_type.py

@@ -1,164 +0,0 @@
-"""Rename entity_type column to note_type
-
-Revision ID: j3d4e5f6g7h8
-Revises: i2c3d4e5f6g7
-Create Date: 2026-02-22 12:00:00.000000
-
-"""
-
-from typing import Sequence, Union
-
-from alembic import op
-from sqlalchemy import text
-
-# revision identifiers, used by Alembic.
-revision: str = "j3d4e5f6g7h8"
-down_revision: Union[str, None] = "i2c3d4e5f6g7"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def table_exists(connection, table_name: str) -> bool:
-    """Check if a table exists (idempotent migration support)."""
-    if connection.dialect.name == "postgresql":
-        result = connection.execute(
-            text("SELECT 1 FROM information_schema.tables WHERE table_name = :table_name"),
-            {"table_name": table_name},
-        )
-        return result.fetchone() is not None
-    # SQLite
-    result = connection.execute(
-        text("SELECT 1 FROM sqlite_master WHERE type='table' AND name = :table_name"),
-        {"table_name": table_name},
-    )
-    return result.fetchone() is not None
-
-
-def index_exists(connection, index_name: str) -> bool:
-    """Check if an index exists (idempotent migration support)."""
-    if connection.dialect.name == "postgresql":
-        result = connection.execute(
-            text("SELECT 1 FROM pg_indexes WHERE indexname = :index_name"),
-            {"index_name": index_name},
-        )
-        return result.fetchone() is not None
-    # SQLite
-    result = connection.execute(
-        text("SELECT 1 FROM sqlite_master WHERE type='index' AND name = :index_name"),
-        {"index_name": index_name},
-    )
-    return result.fetchone() is not None
-
-
-def column_exists(connection, table: str, column: str) -> bool:
-    """Check if a column exists in a table (idempotent migration support)."""
-    if connection.dialect.name == "postgresql":
-        result = connection.execute(
-            text(
-                "SELECT 1 FROM information_schema.columns "
-                "WHERE table_name = :table AND column_name = :column"
-            ),
-            {"table": table, "column": column},
-        )
-        return result.fetchone() is not None
-    # SQLite
-    result = connection.execute(text(f"PRAGMA table_info({table})"))
-    columns = [row[1] for row in result]
-    return column in columns
-
-
-def upgrade() -> None:
-    """Rename entity_type → note_type on the entity table."""
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    # Skip if already migrated (idempotent)
-    if column_exists(connection, "entity", "note_type"):
-        return
-
-    if dialect == "postgresql":
-        # Postgres supports direct column rename
-        op.execute("ALTER TABLE entity RENAME COLUMN entity_type TO note_type")
-
-        # Recreate the index with new name
-        op.execute("DROP INDEX IF EXISTS ix_entity_type")
-        op.execute("CREATE INDEX ix_note_type ON entity (note_type)")
-    else:
-        # SQLite 3.25.0+ supports ALTER TABLE RENAME COLUMN directly.
-        # Avoids batch_alter_table which fails on tables with generated columns
-        # (duplicate column name error when recreating the table).
-        op.execute("ALTER TABLE entity RENAME COLUMN entity_type TO note_type")
-
-        # Recreate the index with new name
-        if index_exists(connection, "ix_entity_type"):
-            op.drop_index("ix_entity_type", table_name="entity")
-        op.create_index("ix_note_type", "entity", ["note_type"])
-
-    # Update search index metadata: rename entity_type → note_type in JSON
-    # This updates the stored metadata so search results use the new field name
-    # Guard: search_index may not exist on a fresh DB (created by an earlier migration)
-    if not table_exists(connection, "search_index"):
-        return
-
-    if dialect == "postgresql":
-        op.execute(
-            text("""
-                UPDATE search_index
-                SET metadata = metadata - 'entity_type' || jsonb_build_object('note_type', metadata->'entity_type')
-                WHERE metadata ? 'entity_type'
-            """)
-        )
-    else:
-        op.execute(
-            text("""
-                UPDATE search_index
-                SET metadata = json_set(
-                    json_remove(metadata, '$.entity_type'),
-                    '$.note_type',
-                    json_extract(metadata, '$.entity_type')
-                )
-                WHERE json_extract(metadata, '$.entity_type') IS NOT NULL
-            """)
-        )
-
-
-def downgrade() -> None:
-    """Rename note_type → entity_type on the entity table."""
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    if dialect == "postgresql":
-        op.execute("ALTER TABLE entity RENAME COLUMN note_type TO entity_type")
-        op.execute("DROP INDEX IF EXISTS ix_note_type")
-        op.execute("CREATE INDEX ix_entity_type ON entity (entity_type)")
-    else:
-        op.execute("ALTER TABLE entity RENAME COLUMN note_type TO entity_type")
-
-        if index_exists(connection, "ix_note_type"):
-            op.drop_index("ix_note_type", table_name="entity")
-        op.create_index("ix_entity_type", "entity", ["entity_type"])
-
-    # Revert search index metadata
-    if not table_exists(connection, "search_index"):
-        return
-
-    if dialect == "postgresql":
-        op.execute(
-            text("""
-                UPDATE search_index
-                SET metadata = metadata - 'note_type' || jsonb_build_object('entity_type', metadata->'note_type')
-                WHERE metadata ? 'note_type'
-            """)
-        )
-    else:
-        op.execute(
-            text("""
-                UPDATE search_index
-                SET metadata = json_set(
-                    json_remove(metadata, '$.note_type'),
-                    '$.entity_type',
-                    json_extract(metadata, '$.note_type')
-                )
-                WHERE json_extract(metadata, '$.note_type') IS NOT NULL
-            """)
-        )

src/basic_memory/alembic/versions/k4e5f6g7h8i9_add_created_by_and_last_updated_by_to_entity.py

@@ -1,74 +0,0 @@
-"""Add created_by and last_updated_by columns to entity table.
-
-Revision ID: k4e5f6g7h8i9
-Revises: j3d4e5f6g7h8
-Create Date: 2026-02-23 00:00:00.000000
-
-These columns track which cloud user created and last modified each entity.
-Both are nullable — NULL for local/CLI usage and existing entities.
-"""
-
-from typing import Sequence, Union
-
-import sqlalchemy as sa
-from alembic import op
-from sqlalchemy import text
-
-# revision identifiers, used by Alembic.
-revision: str = "k4e5f6g7h8i9"
-down_revision: Union[str, None] = "j3d4e5f6g7h8"
-branch_labels: Union[str, Sequence[str], None] = None
-depends_on: Union[str, Sequence[str], None] = None
-
-
-def column_exists(connection, table: str, column: str) -> bool:
-    """Check if a column exists in a table (idempotent migration support)."""
-    if connection.dialect.name == "postgresql":
-        result = connection.execute(
-            text(
-                "SELECT 1 FROM information_schema.columns "
-                "WHERE table_name = :table AND column_name = :column"
-            ),
-            {"table": table, "column": column},
-        )
-        return result.fetchone() is not None
-    else:
-        # SQLite
-        result = connection.execute(text(f"PRAGMA table_info({table})"))
-        columns = [row[1] for row in result]
-        return column in columns
-
-
-def upgrade() -> None:
-    """Add created_by and last_updated_by columns to entity table.
-
-    Both columns are nullable strings that store cloud user_profile_id UUIDs.
-    No data backfill — existing rows get NULL.
-    """
-    connection = op.get_bind()
-
-    if not column_exists(connection, "entity", "created_by"):
-        op.add_column("entity", sa.Column("created_by", sa.String(), nullable=True))
-
-    if not column_exists(connection, "entity", "last_updated_by"):
-        op.add_column("entity", sa.Column("last_updated_by", sa.String(), nullable=True))
-
-
-def downgrade() -> None:
-    """Remove created_by and last_updated_by columns from entity table."""
-    connection = op.get_bind()
-    dialect = connection.dialect.name
-
-    if column_exists(connection, "entity", "last_updated_by"):
-        if dialect == "postgresql":
-            op.drop_column("entity", "last_updated_by")
-        else:
-            with op.batch_alter_table("entity") as batch_op:
-                batch_op.drop_column("last_updated_by")
-
-    if column_exists(connection, "entity", "created_by"):
-        if dialect == "postgresql":
-            op.drop_column("entity", "created_by")
-        else:
-            with op.batch_alter_table("entity") as batch_op:
-                batch_op.drop_column("created_by")

src/basic_memory/api/app.py

@@ -1,72 +1,52 @@
 """FastAPI application for basic-memory knowledge graph API."""
 
+import asyncio
 from contextlib import asynccontextmanager
 
-from fastapi import FastAPI, HTTPException, Request
+from fastapi import FastAPI, HTTPException
 from fastapi.exception_handlers import http_exception_handler
-from fastapi.routing import APIRouter
 from loguru import logger
 
 from basic_memory import __version__ as version
-from basic_memory.api.container import ApiContainer, set_container
-from basic_memory.api.v2.routers import (
-    knowledge_router as v2_knowledge,
-    project_router as v2_project,
-    memory_router as v2_memory,
-    search_router as v2_search,
-    resource_router as v2_resource,
-    directory_router as v2_directory,
-    prompt_router as v2_prompt,
-    importer_router as v2_importer,
-    schema_router as v2_schema,
+from basic_memory import db
+from basic_memory.api.routers import (
+    directory_router,
+    importer_router,
+    knowledge,
+    management,
+    memory,
+    project,
+    resource,
+    search,
+    prompt_router,
 )
-from basic_memory.api.v2.routers.project_router import (
-    add_project,
-    list_projects,
-    synchronize_projects,
-)
-from basic_memory.config import init_api_logging
-from basic_memory.services.exceptions import EntityAlreadyExistsError
-from basic_memory.services.initialization import initialize_app
+from basic_memory.config import app_config
+from basic_memory.services.initialization import initialize_app, initialize_file_sync
 
 
 @asynccontextmanager
 async def lifespan(app: FastAPI):  # pragma: no cover
-    """Lifecycle manager for the FastAPI app. Not called in stdio mcp mode"""
+    """Lifecycle manager for the FastAPI app."""
+    # Initialize app and database
+    logger.info("Starting Basic Memory API")
+    await initialize_app(app_config)
 
-    # Initialize logging for API (stdout in cloud mode, file otherwise)
-    init_api_logging()
+    logger.info(f"Sync changes enabled: {app_config.sync_changes}")
+    if app_config.sync_changes:
+        # start file sync task in background
+        app.state.sync_task = asyncio.create_task(initialize_file_sync(app_config))
+    else:
+        logger.info("Sync changes disabled. Skipping file sync service.")
 
-    # --- Composition Root ---
-    # Create container and read config (single point of config access)
-    container = ApiContainer.create()
-    set_container(container)
-    app.state.container = container
-
-    logger.info(f"Starting Basic Memory API (mode={container.mode.name})")
-
-    await initialize_app(container.config)
-
-    # Cache database connections in app state for performance
-    logger.info("Initializing database and caching connections...")
-    engine, session_maker = await container.init_database()
-    app.state.engine = engine
-    app.state.session_maker = session_maker
-    logger.info("Database connections cached in app state")
-
-    # Create and start sync coordinator (lifecycle centralized in coordinator)
-    sync_coordinator = container.create_sync_coordinator()
-    await sync_coordinator.start()
-    app.state.sync_coordinator = sync_coordinator
-
-    # Proceed with startup
+    # proceed with startup
     yield
 
-    # Shutdown - coordinator handles clean task cancellation
     logger.info("Shutting down Basic Memory API")
-    await sync_coordinator.stop()
+    if app.state.sync_task:
+        logger.info("Stopping sync...")
+        app.state.sync_task.cancel()  # pyright: ignore
 
-    await container.shutdown_database()
+    await db.shutdown_db()
 
 
 # Initialize FastAPI app
@@ -77,52 +57,22 @@ app = FastAPI(
     lifespan=lifespan,
 )
 
-# Include v2 routers FIRST (more specific paths must match before /{project} catch-all)
-app.include_router(v2_knowledge, prefix="/v2/projects/{project_id}")
-app.include_router(v2_memory, prefix="/v2/projects/{project_id}")
-app.include_router(v2_search, prefix="/v2/projects/{project_id}")
-app.include_router(v2_resource, prefix="/v2/projects/{project_id}")
-app.include_router(v2_directory, prefix="/v2/projects/{project_id}")
-app.include_router(v2_prompt, prefix="/v2/projects/{project_id}")
-app.include_router(v2_importer, prefix="/v2/projects/{project_id}")
-app.include_router(v2_schema, prefix="/v2/projects/{project_id}")
-app.include_router(v2_project, prefix="/v2")
 
-# Legacy web app proxy paths (compat with /proxy/projects/projects)
-app.include_router(v2_project, prefix="/proxy/projects")
+# Include routers
+app.include_router(knowledge.router, prefix="/{project}")
+app.include_router(memory.router, prefix="/{project}")
+app.include_router(resource.router, prefix="/{project}")
+app.include_router(search.router, prefix="/{project}")
+app.include_router(project.project_router, prefix="/{project}")
+app.include_router(directory_router.router, prefix="/{project}")
+app.include_router(prompt_router.router, prefix="/{project}")
+app.include_router(importer_router.router, prefix="/{project}")
 
-# Legacy v1 compat: older CLI versions (v0.18.0 and earlier) call /projects/...
-# Using router mount causes 307 redirect which proxy doesn't follow, so add explicit routes
-legacy_router = APIRouter(tags=["legacy"])
-legacy_router.add_api_route("/projects/projects", list_projects, methods=["GET"])
-legacy_router.add_api_route("/projects/projects", add_project, methods=["POST"])
-legacy_router.add_api_route("/projects/config/sync", synchronize_projects, methods=["POST"])
-app.include_router(legacy_router)
+# Project resource router works accross projects
+app.include_router(project.project_resource_router)
+app.include_router(management.router)
 
-# V2 routers are the only public API surface
-
-
-@app.exception_handler(EntityAlreadyExistsError)
-async def entity_already_exists_error_handler(request: Request, exc: EntityAlreadyExistsError):
-    """Handle entity creation conflicts (e.g., file already exists).
-
-    This is expected behavior when users try to create notes that exist,
-    so log at INFO level instead of ERROR.
-    """
-    logger.info(
-        "Entity already exists",
-        url=str(request.url),
-        method=request.method,
-        path=request.url.path,
-        error=str(exc),
-    )
-    return await http_exception_handler(
-        request,
-        HTTPException(
-            status_code=409,
-            detail="Note already exists. Use edit_note to modify it, or delete it first.",
-        ),
-    )
+# Auth routes are handled by FastMCP automatically when auth is enabled
 
 
 @app.exception_handler(Exception)

src/basic_memory/api/container.py

@@ -1,132 +0,0 @@
-"""API composition root for Basic Memory.
-
-This container owns reading ConfigManager and environment variables for the
-API entrypoint. Downstream modules receive config/dependencies explicitly
-rather than reading globals.
-
-Design principles:
-- Only this module reads ConfigManager directly
-- Runtime mode (cloud/local/test) is resolved here
-- Factories for services are provided, not singletons
-"""
-
-from dataclasses import dataclass
-from typing import TYPE_CHECKING
-
-from sqlalchemy.ext.asyncio import AsyncEngine, async_sessionmaker, AsyncSession
-
-from basic_memory import db
-from basic_memory.config import BasicMemoryConfig, ConfigManager
-from basic_memory.runtime import RuntimeMode, resolve_runtime_mode
-
-if TYPE_CHECKING:  # pragma: no cover
-    from basic_memory.sync import SyncCoordinator
-
-
-@dataclass
-class ApiContainer:
-    """Composition root for the API entrypoint.
-
-    Holds resolved configuration and runtime context.
-    Created once at app startup, then used to wire dependencies.
-    """
-
-    config: BasicMemoryConfig
-    mode: RuntimeMode
-
-    # --- Database ---
-    # Cached database connections (set during lifespan startup)
-    engine: AsyncEngine | None = None
-    session_maker: async_sessionmaker[AsyncSession] | None = None
-
-    @classmethod
-    def create(cls) -> "ApiContainer":  # pragma: no cover
-        """Create container by reading ConfigManager.
-
-        This is the single point where API reads global config.
-        """
-        config = ConfigManager().config
-        mode = resolve_runtime_mode(
-            is_test_env=config.is_test_env,
-        )
-        return cls(config=config, mode=mode)
-
-    # --- Runtime Mode Properties ---
-
-    @property
-    def should_sync_files(self) -> bool:
-        """Whether file sync should be started.
-
-        Sync is enabled when:
-        - sync_changes is True in config
-        - Not in test mode (tests manage their own sync)
-        """
-        return self.config.sync_changes and not self.mode.is_test
-
-    @property
-    def sync_skip_reason(self) -> str | None:  # pragma: no cover
-        """Reason why sync is skipped, or None if sync should run.
-
-        Useful for logging why sync was disabled.
-        """
-        if self.mode.is_test:
-            return "Test environment detected"
-        if not self.config.sync_changes:
-            return "Sync changes disabled"
-        return None
-
-    def create_sync_coordinator(self) -> "SyncCoordinator":  # pragma: no cover
-        """Create a SyncCoordinator with this container's settings.
-
-        Returns:
-            SyncCoordinator configured for this runtime environment
-        """
-        # Deferred import to avoid circular dependency
-        from basic_memory.sync import SyncCoordinator
-
-        return SyncCoordinator(
-            config=self.config,
-            should_sync=self.should_sync_files,
-            skip_reason=self.sync_skip_reason,
-        )
-
-    # --- Database Factory ---
-
-    async def init_database(  # pragma: no cover
-        self,
-    ) -> tuple[AsyncEngine, async_sessionmaker[AsyncSession]]:
-        """Initialize and cache database connections.
-
-        Returns:
-            Tuple of (engine, session_maker)
-        """
-        engine, session_maker = await db.get_or_create_db(self.config.database_path)
-        self.engine = engine
-        self.session_maker = session_maker
-        return engine, session_maker
-
-    async def shutdown_database(self) -> None:  # pragma: no cover
-        """Clean up database connections."""
-        await db.shutdown_db()
-
-
-# Module-level container instance (set by lifespan)
-# This allows deps.py to access the container without reading ConfigManager
-_container: ApiContainer | None = None
-
-
-def get_container() -> ApiContainer:
-    """Get the current API container.
-
-    Raises:
-        RuntimeError: If container hasn't been initialized
-    """
-    if _container is None:
-        raise RuntimeError("API container not initialized. Call set_container() first.")
-    return _container
-
-
-def set_container(container: ApiContainer) -> None:
-    """Set the API container (called by lifespan)."""
-    global _container
-    _container = container

src/basic_memory/api/routers/__init__.py

@@ -0,0 +1,11 @@
+"""API routers."""
+
+from . import knowledge_router as knowledge
+from . import management_router as management
+from . import memory_router as memory
+from . import project_router as project
+from . import resource_router as resource
+from . import search_router as search
+from . import prompt_router as prompt
+
+__all__ = ["knowledge", "management", "memory", "project", "resource", "search", "prompt"]

src/basic_memory/api/routers/directory_router.py

@@ -0,0 +1,63 @@
+"""Router for directory tree operations."""
+
+from typing import List, Optional
+
+from fastapi import APIRouter, Query
+
+from basic_memory.deps import DirectoryServiceDep, ProjectIdDep
+from basic_memory.schemas.directory import DirectoryNode
+
+router = APIRouter(prefix="/directory", tags=["directory"])
+
+
+@router.get("/tree", response_model=DirectoryNode)
+async def get_directory_tree(
+    directory_service: DirectoryServiceDep,
+    project_id: ProjectIdDep,
+):
+    """Get hierarchical directory structure from the knowledge base.
+
+    Args:
+        directory_service: Service for directory operations
+        project_id: ID of the current project
+
+    Returns:
+        DirectoryNode representing the root of the hierarchical tree structure
+    """
+    # Get a hierarchical directory tree for the specific project
+    tree = await directory_service.get_directory_tree()
+
+    # Return the hierarchical tree
+    return tree
+
+
+@router.get("/list", response_model=List[DirectoryNode])
+async def list_directory(
+    directory_service: DirectoryServiceDep,
+    project_id: ProjectIdDep,
+    dir_name: str = Query("/", description="Directory path to list"),
+    depth: int = Query(1, ge=1, le=10, description="Recursion depth (1-10)"),
+    file_name_glob: Optional[str] = Query(
+        None, description="Glob pattern for filtering file names"
+    ),
+):
+    """List directory contents with filtering and depth control.
+
+    Args:
+        directory_service: Service for directory operations
+        project_id: ID of the current project
+        dir_name: Directory path to list (default: root "/")
+        depth: Recursion depth (1-10, default: 1 for immediate children only)
+        file_name_glob: Optional glob pattern for filtering file names (e.g., "*.md", "*meeting*")
+
+    Returns:
+        List of DirectoryNode objects matching the criteria
+    """
+    # Get directory listing with filtering
+    nodes = await directory_service.list_directory(
+        dir_name=dir_name,
+        depth=depth,
+        file_name_glob=file_name_glob,
+    )
+
+    return nodes

src/basic_memory/api/routers/importer_router.py

@@ -1,19 +1,15 @@
-"""V2 Import Router - ID-based data import operations.
-
-This router uses v2 dependencies for consistent project handling with external_id UUIDs.
-Import endpoints use project_id in the path for consistency with other v2 endpoints.
-"""
+"""Import router for Basic Memory API."""
 
 import json
 import logging
 
-from fastapi import APIRouter, Form, HTTPException, UploadFile, status, Path
+from fastapi import APIRouter, Form, HTTPException, UploadFile, status
 
 from basic_memory.deps import (
-    ChatGPTImporterV2ExternalDep,
-    ClaudeConversationsImporterV2ExternalDep,
-    ClaudeProjectsImporterV2ExternalDep,
-    MemoryJsonImporterV2ExternalDep,
+    ChatGPTImporterDep,
+    ClaudeConversationsImporterDep,
+    ClaudeProjectsImporterDep,
+    MemoryJsonImporterDep,
 )
 from basic_memory.importers import Importer
 from basic_memory.schemas.importer import (
@@ -24,23 +20,21 @@ from basic_memory.schemas.importer import (
 
 logger = logging.getLogger(__name__)
 
-router = APIRouter(prefix="/import", tags=["import-v2"])
+router = APIRouter(prefix="/import", tags=["import"])
 
 
 @router.post("/chatgpt", response_model=ChatImportResult)
 async def import_chatgpt(
-    importer: ChatGPTImporterV2ExternalDep,
+    importer: ChatGPTImporterDep,
     file: UploadFile,
-    project_id: str = Path(..., description="Project external UUID"),
-    directory: str = Form("conversations"),
+    folder: str = Form("conversations"),
 ) -> ChatImportResult:
     """Import conversations from ChatGPT JSON export.
 
     Args:
-        project_id: Project external UUID from URL path
         file: The ChatGPT conversations.json file.
-        directory: The directory to place the files in.
-        importer: ChatGPT importer instance.
+        folder: The folder to place the files in.
+        markdown_processor: MarkdownProcessor instance.
 
     Returns:
         ChatImportResult with import statistics.
@@ -48,24 +42,21 @@ async def import_chatgpt(
     Raises:
         HTTPException: If import fails.
     """
-    logger.info(f"V2 Importing ChatGPT conversations for project {project_id}")
-    return await import_file(importer, file, directory)
+    return await import_file(importer, file, folder)
 
 
 @router.post("/claude/conversations", response_model=ChatImportResult)
 async def import_claude_conversations(
-    importer: ClaudeConversationsImporterV2ExternalDep,
+    importer: ClaudeConversationsImporterDep,
     file: UploadFile,
-    project_id: str = Path(..., description="Project external UUID"),
-    directory: str = Form("conversations"),
+    folder: str = Form("conversations"),
 ) -> ChatImportResult:
     """Import conversations from Claude conversations.json export.
 
     Args:
-        project_id: Project external UUID from URL path
         file: The Claude conversations.json file.
-        directory: The directory to place the files in.
-        importer: Claude conversations importer instance.
+        folder: The folder to place the files in.
+        markdown_processor: MarkdownProcessor instance.
 
     Returns:
         ChatImportResult with import statistics.
@@ -73,24 +64,21 @@ async def import_claude_conversations(
     Raises:
         HTTPException: If import fails.
     """
-    logger.info(f"V2 Importing Claude conversations for project {project_id}")
-    return await import_file(importer, file, directory)
+    return await import_file(importer, file, folder)
 
 
 @router.post("/claude/projects", response_model=ProjectImportResult)
 async def import_claude_projects(
-    importer: ClaudeProjectsImporterV2ExternalDep,
+    importer: ClaudeProjectsImporterDep,
     file: UploadFile,
-    project_id: str = Path(..., description="Project external UUID"),
-    directory: str = Form("projects"),
+    folder: str = Form("projects"),
 ) -> ProjectImportResult:
     """Import projects from Claude projects.json export.
 
     Args:
-        project_id: Project external UUID from URL path
         file: The Claude projects.json file.
-        directory: The base directory to place the files in.
-        importer: Claude projects importer instance.
+        base_folder: The base folder to place the files in.
+        markdown_processor: MarkdownProcessor instance.
 
     Returns:
         ProjectImportResult with import statistics.
@@ -98,24 +86,21 @@ async def import_claude_projects(
     Raises:
         HTTPException: If import fails.
     """
-    logger.info(f"V2 Importing Claude projects for project {project_id}")
-    return await import_file(importer, file, directory)
+    return await import_file(importer, file, folder)
 
 
 @router.post("/memory-json", response_model=EntityImportResult)
 async def import_memory_json(
-    importer: MemoryJsonImporterV2ExternalDep,
+    importer: MemoryJsonImporterDep,
     file: UploadFile,
-    project_id: str = Path(..., description="Project external UUID"),
-    directory: str = Form("conversations"),
+    folder: str = Form("conversations"),
 ) -> EntityImportResult:
     """Import entities and relations from a memory.json file.
 
     Args:
-        project_id: Project external UUID from URL path
         file: The memory.json file.
-        directory: Optional destination directory within the project.
-        importer: Memory JSON importer instance.
+        destination_folder: Optional destination folder within the project.
+        markdown_processor: MarkdownProcessor instance.
 
     Returns:
         EntityImportResult with import statistics.
@@ -123,7 +108,6 @@ async def import_memory_json(
     Raises:
         HTTPException: If import fails.
     """
-    logger.info(f"V2 Importing memory.json for project {project_id}")
     try:
         file_data = []
         file_bytes = await file.read()
@@ -132,14 +116,14 @@ async def import_memory_json(
             json_data = json.loads(line)
             file_data.append(json_data)
 
-        result = await importer.import_data(file_data, directory)
+        result = await importer.import_data(file_data, folder)
         if not result.success:  # pragma: no cover
             raise HTTPException(
                 status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
                 detail=result.error_message or "Import failed",
             )
     except Exception as e:
-        logger.exception("V2 Import failed")
+        logger.exception("Import failed")
         raise HTTPException(
             status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
             detail=f"Import failed: {str(e)}",
@@ -147,24 +131,11 @@ async def import_memory_json(
     return result
 
 
-async def import_file(importer: Importer, file: UploadFile, destination_directory: str):
-    """Helper function to import a file using an importer instance.
-
-    Args:
-        importer: The importer instance to use
-        file: The file to import
-        destination_directory: Destination directory for imported content
-
-    Returns:
-        Import result from the importer
-
-    Raises:
-        HTTPException: If import fails
-    """
+async def import_file(importer: Importer, file: UploadFile, destination_folder: str):
     try:
         # Process file
         json_data = json.load(file.file)
-        result = await importer.import_data(json_data, destination_directory)
+        result = await importer.import_data(json_data, destination_folder)
         if not result.success:  # pragma: no cover
             raise HTTPException(
                 status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
@@ -174,7 +145,7 @@ async def import_file(importer: Importer, file: UploadFile, destination_director
         return result
 
     except Exception as e:
-        logger.exception("V2 Import failed")
+        logger.exception("Import failed")
         raise HTTPException(
             status_code=status.HTTP_500_INTERNAL_SERVER_ERROR,
             detail=f"Import failed: {str(e)}",

src/basic_memory/api/routers/knowledge_router.py

@@ -0,0 +1,290 @@
+"""Router for knowledge graph operations."""
+
+from typing import Annotated
+
+from fastapi import APIRouter, HTTPException, BackgroundTasks, Depends, Query, Response
+from loguru import logger
+
+from basic_memory.deps import (
+    EntityServiceDep,
+    get_search_service,
+    SearchServiceDep,
+    LinkResolverDep,
+    ProjectPathDep,
+    FileServiceDep,
+    ProjectConfigDep,
+    AppConfigDep,
+    SyncServiceDep,
+)
+from basic_memory.schemas import (
+    EntityListResponse,
+    EntityResponse,
+    DeleteEntitiesResponse,
+    DeleteEntitiesRequest,
+)
+from basic_memory.schemas.request import EditEntityRequest, MoveEntityRequest
+from basic_memory.schemas.base import Permalink, Entity
+
+router = APIRouter(prefix="/knowledge", tags=["knowledge"])
+
+## Create endpoints
+
+
+@router.post("/entities", response_model=EntityResponse)
+async def create_entity(
+    data: Entity,
+    background_tasks: BackgroundTasks,
+    entity_service: EntityServiceDep,
+    search_service: SearchServiceDep,
+) -> EntityResponse:
+    """Create an entity."""
+    logger.info(
+        "API request", endpoint="create_entity", entity_type=data.entity_type, title=data.title
+    )
+
+    entity = await entity_service.create_entity(data)
+
+    # reindex
+    await search_service.index_entity(entity, background_tasks=background_tasks)
+    result = EntityResponse.model_validate(entity)
+
+    logger.info(
+        f"API response: endpoint='create_entity' title={result.title}, permalink={result.permalink}, status_code=201"
+    )
+    return result
+
+
+@router.put("/entities/{permalink:path}", response_model=EntityResponse)
+async def create_or_update_entity(
+    project: ProjectPathDep,
+    permalink: Permalink,
+    data: Entity,
+    response: Response,
+    background_tasks: BackgroundTasks,
+    entity_service: EntityServiceDep,
+    search_service: SearchServiceDep,
+    file_service: FileServiceDep,
+    sync_service: SyncServiceDep,
+) -> EntityResponse:
+    """Create or update an entity. If entity exists, it will be updated, otherwise created."""
+    logger.info(
+        f"API request: create_or_update_entity for {project=}, {permalink=}, {data.entity_type=}, {data.title=}"
+    )
+
+    # Validate permalink matches
+    if data.permalink != permalink:
+        logger.warning(
+            f"API validation error: creating/updating entity with permalink mismatch - url={permalink}, data={data.permalink}",
+        )
+        raise HTTPException(
+            status_code=400,
+            detail=f"Entity permalink {data.permalink} must match URL path: '{permalink}'",
+        )
+
+    # Try create_or_update operation
+    entity, created = await entity_service.create_or_update_entity(data)
+    response.status_code = 201 if created else 200
+
+    # reindex
+    await search_service.index_entity(entity, background_tasks=background_tasks)
+
+    # Attempt immediate relation resolution when creating new entities
+    # This helps resolve forward references when related entities are created in the same session
+    if created:
+        try:
+            await sync_service.resolve_relations()
+            logger.debug(f"Resolved relations after creating entity: {entity.permalink}")
+        except Exception as e:  # pragma: no cover
+            # Don't fail the entire request if relation resolution fails
+            logger.warning(f"Failed to resolve relations after entity creation: {e}")
+
+    result = EntityResponse.model_validate(entity)
+
+    logger.info(
+        f"API response: {result.title=}, {result.permalink=}, {created=}, status_code={response.status_code}"
+    )
+    return result
+
+
+@router.patch("/entities/{identifier:path}", response_model=EntityResponse)
+async def edit_entity(
+    identifier: str,
+    data: EditEntityRequest,
+    background_tasks: BackgroundTasks,
+    entity_service: EntityServiceDep,
+    search_service: SearchServiceDep,
+) -> EntityResponse:
+    """Edit an existing entity using various operations like append, prepend, find_replace, or replace_section.
+
+    This endpoint allows for targeted edits without requiring the full entity content.
+    """
+    logger.info(
+        f"API request: endpoint='edit_entity', identifier='{identifier}', operation='{data.operation}'"
+    )
+
+    try:
+        # Edit the entity using the service
+        entity = await entity_service.edit_entity(
+            identifier=identifier,
+            operation=data.operation,
+            content=data.content,
+            section=data.section,
+            find_text=data.find_text,
+            expected_replacements=data.expected_replacements,
+        )
+
+        # Reindex the updated entity
+        await search_service.index_entity(entity, background_tasks=background_tasks)
+
+        # Return the updated entity response
+        result = EntityResponse.model_validate(entity)
+
+        logger.info(
+            "API response",
+            endpoint="edit_entity",
+            identifier=identifier,
+            operation=data.operation,
+            permalink=result.permalink,
+            status_code=200,
+        )
+
+        return result
+
+    except Exception as e:
+        logger.error(f"Error editing entity: {e}")
+        raise HTTPException(status_code=400, detail=str(e))
+
+
+@router.post("/move")
+async def move_entity(
+    data: MoveEntityRequest,
+    background_tasks: BackgroundTasks,
+    entity_service: EntityServiceDep,
+    project_config: ProjectConfigDep,
+    app_config: AppConfigDep,
+    search_service: SearchServiceDep,
+) -> EntityResponse:
+    """Move an entity to a new file location with project consistency.
+
+    This endpoint moves a note to a different path while maintaining project
+    consistency and optionally updating permalinks based on configuration.
+    """
+    logger.info(
+        f"API request: endpoint='move_entity', identifier='{data.identifier}', destination='{data.destination_path}'"
+    )
+
+    try:
+        # Move the entity using the service
+        moved_entity = await entity_service.move_entity(
+            identifier=data.identifier,
+            destination_path=data.destination_path,
+            project_config=project_config,
+            app_config=app_config,
+        )
+
+        # Get the moved entity to reindex it
+        entity = await entity_service.link_resolver.resolve_link(data.destination_path)
+        if entity:
+            await search_service.index_entity(entity, background_tasks=background_tasks)
+
+        logger.info(
+            "API response",
+            endpoint="move_entity",
+            identifier=data.identifier,
+            destination=data.destination_path,
+            status_code=200,
+        )
+        result = EntityResponse.model_validate(moved_entity)
+        return result
+
+    except Exception as e:
+        logger.error(f"Error moving entity: {e}")
+        raise HTTPException(status_code=400, detail=str(e))
+
+
+## Read endpoints
+
+
+@router.get("/entities/{identifier:path}", response_model=EntityResponse)
+async def get_entity(
+    entity_service: EntityServiceDep,
+    link_resolver: LinkResolverDep,
+    identifier: str,
+) -> EntityResponse:
+    """Get a specific entity by file path or permalink..
+
+    Args:
+        identifier: Entity file path or permalink
+        :param entity_service: EntityService
+        :param link_resolver: LinkResolver
+    """
+    logger.info(f"request: get_entity with identifier={identifier}")
+    entity = await link_resolver.resolve_link(identifier)
+    if not entity:
+        raise HTTPException(status_code=404, detail=f"Entity {identifier} not found")
+
+    result = EntityResponse.model_validate(entity)
+    return result
+
+
+@router.get("/entities", response_model=EntityListResponse)
+async def get_entities(
+    entity_service: EntityServiceDep,
+    permalink: Annotated[list[str] | None, Query()] = None,
+) -> EntityListResponse:
+    """Open specific entities"""
+    logger.info(f"request: get_entities with permalinks={permalink}")
+
+    entities = await entity_service.get_entities_by_permalinks(permalink) if permalink else []
+    result = EntityListResponse(
+        entities=[EntityResponse.model_validate(entity) for entity in entities]
+    )
+    return result
+
+
+## Delete endpoints
+
+
+@router.delete("/entities/{identifier:path}", response_model=DeleteEntitiesResponse)
+async def delete_entity(
+    identifier: str,
+    background_tasks: BackgroundTasks,
+    entity_service: EntityServiceDep,
+    link_resolver: LinkResolverDep,
+    search_service=Depends(get_search_service),
+) -> DeleteEntitiesResponse:
+    """Delete a single entity and remove from search index."""
+    logger.info(f"request: delete_entity with identifier={identifier}")
+
+    entity = await link_resolver.resolve_link(identifier)
+    if entity is None:
+        return DeleteEntitiesResponse(deleted=False)
+
+    # Delete the entity
+    deleted = await entity_service.delete_entity(entity.permalink or entity.id)
+
+    # Remove from search index (entity, observations, and relations)
+    background_tasks.add_task(search_service.handle_delete, entity)
+
+    result = DeleteEntitiesResponse(deleted=deleted)
+    return result
+
+
+@router.post("/entities/delete", response_model=DeleteEntitiesResponse)
+async def delete_entities(
+    data: DeleteEntitiesRequest,
+    background_tasks: BackgroundTasks,
+    entity_service: EntityServiceDep,
+    search_service=Depends(get_search_service),
+) -> DeleteEntitiesResponse:
+    """Delete entities and remove from search index."""
+    logger.info(f"request: delete_entities with data={data}")
+    deleted = False
+
+    # Remove each deleted entity from search index
+    for permalink in data.permalinks:
+        deleted = await entity_service.delete_entity(permalink)
+        background_tasks.add_task(search_service.delete_by_permalink, permalink)
+
+    result = DeleteEntitiesResponse(deleted=deleted)
+    return result

src/basic_memory/api/routers/management_router.py

@@ -0,0 +1,78 @@
+"""Management router for basic-memory API."""
+
+import asyncio
+
+from fastapi import APIRouter, Request
+from loguru import logger
+from pydantic import BaseModel
+
+from basic_memory.config import app_config
+from basic_memory.deps import SyncServiceDep, ProjectRepositoryDep
+
+router = APIRouter(prefix="/management", tags=["management"])
+
+
+class WatchStatusResponse(BaseModel):
+    """Response model for watch status."""
+
+    running: bool
+    """Whether the watch service is currently running."""
+
+
+@router.get("/watch/status", response_model=WatchStatusResponse)
+async def get_watch_status(request: Request) -> WatchStatusResponse:
+    """Get the current status of the watch service."""
+    return WatchStatusResponse(
+        running=request.app.state.watch_task is not None and not request.app.state.watch_task.done()
+    )
+
+
+@router.post("/watch/start", response_model=WatchStatusResponse)
+async def start_watch_service(
+    request: Request, project_repository: ProjectRepositoryDep, sync_service: SyncServiceDep
+) -> WatchStatusResponse:
+    """Start the watch service if it's not already running."""
+
+    # needed because of circular imports from sync -> app
+    from basic_memory.sync import WatchService
+    from basic_memory.sync.background_sync import create_background_sync_task
+
+    if request.app.state.watch_task is not None and not request.app.state.watch_task.done():
+        # Watch service is already running
+        return WatchStatusResponse(running=True)
+
+    # Create and start a new watch service
+    logger.info("Starting watch service via management API")
+
+    # Get services needed for the watch task
+    watch_service = WatchService(
+        app_config=app_config,
+        project_repository=project_repository,
+    )
+
+    # Create and store the task
+    watch_task = create_background_sync_task(sync_service, watch_service)
+    request.app.state.watch_task = watch_task
+
+    return WatchStatusResponse(running=True)
+
+
+@router.post("/watch/stop", response_model=WatchStatusResponse)
+async def stop_watch_service(request: Request) -> WatchStatusResponse:  # pragma: no cover
+    """Stop the watch service if it's running."""
+    if request.app.state.watch_task is None or request.app.state.watch_task.done():
+        # Watch service is not running
+        return WatchStatusResponse(running=False)
+
+    # Cancel the running task
+    logger.info("Stopping watch service via management API")
+    request.app.state.watch_task.cancel()
+
+    # Wait for it to be properly cancelled
+    try:
+        await request.app.state.watch_task
+    except asyncio.CancelledError:
+        pass
+
+    request.app.state.watch_task = None
+    return WatchStatusResponse(running=False)

src/basic_memory/api/routers/memory_router.py

@@ -0,0 +1,90 @@
+"""Routes for memory:// URI operations."""
+
+from typing import Annotated, Optional
+
+from fastapi import APIRouter, Query
+from loguru import logger
+
+from basic_memory.deps import ContextServiceDep, EntityRepositoryDep
+from basic_memory.schemas.base import TimeFrame, parse_timeframe
+from basic_memory.schemas.memory import (
+    GraphContext,
+    normalize_memory_url,
+)
+from basic_memory.schemas.search import SearchItemType
+from basic_memory.api.routers.utils import to_graph_context
+
+router = APIRouter(prefix="/memory", tags=["memory"])
+
+
+@router.get("/recent", response_model=GraphContext)
+async def recent(
+    context_service: ContextServiceDep,
+    entity_repository: EntityRepositoryDep,
+    type: Annotated[list[SearchItemType] | None, Query()] = None,
+    depth: int = 1,
+    timeframe: TimeFrame = "7d",
+    page: int = 1,
+    page_size: int = 10,
+    max_related: int = 10,
+) -> GraphContext:
+    # return all types by default
+    types = (
+        [SearchItemType.ENTITY, SearchItemType.RELATION, SearchItemType.OBSERVATION]
+        if not type
+        else type
+    )
+
+    logger.debug(
+        f"Getting recent context: `{types}` depth: `{depth}` timeframe: `{timeframe}` page: `{page}` page_size: `{page_size}` max_related: `{max_related}`"
+    )
+    # Parse timeframe
+    since = parse_timeframe(timeframe)
+    limit = page_size
+    offset = (page - 1) * page_size
+
+    # Build context
+    context = await context_service.build_context(
+        types=types, depth=depth, since=since, limit=limit, offset=offset, max_related=max_related
+    )
+    recent_context = await to_graph_context(
+        context, entity_repository=entity_repository, page=page, page_size=page_size
+    )
+    logger.debug(f"Recent context: {recent_context.model_dump_json()}")
+    return recent_context
+
+
+# get_memory_context needs to be declared last so other paths can match
+
+
+@router.get("/{uri:path}", response_model=GraphContext)
+async def get_memory_context(
+    context_service: ContextServiceDep,
+    entity_repository: EntityRepositoryDep,
+    uri: str,
+    depth: int = 1,
+    timeframe: Optional[TimeFrame] = None,
+    page: int = 1,
+    page_size: int = 10,
+    max_related: int = 10,
+) -> GraphContext:
+    """Get rich context from memory:// URI."""
+    # add the project name from the config to the url as the "host
+    # Parse URI
+    logger.debug(
+        f"Getting context for URI: `{uri}` depth: `{depth}` timeframe: `{timeframe}` page: `{page}` page_size: `{page_size}` max_related: `{max_related}`"
+    )
+    memory_url = normalize_memory_url(uri)
+
+    # Parse timeframe
+    since = parse_timeframe(timeframe) if timeframe else None
+    limit = page_size
+    offset = (page - 1) * page_size
+
+    # Build context
+    context = await context_service.build_context(
+        memory_url, depth=depth, since=since, limit=limit, offset=offset, max_related=max_related
+    )
+    return await to_graph_context(
+        context, entity_repository=entity_repository, page=page, page_size=page_size
+    )

src/basic_memory/api/routers/project_router.py

@@ -0,0 +1,234 @@
+"""Router for project management."""
+
+from fastapi import APIRouter, HTTPException, Path, Body
+from typing import Optional
+
+from basic_memory.deps import ProjectServiceDep, ProjectPathDep
+from basic_memory.schemas import ProjectInfoResponse
+from basic_memory.schemas.project_info import (
+    ProjectList,
+    ProjectItem,
+    ProjectInfoRequest,
+    ProjectStatusResponse,
+)
+
+# Router for resources in a specific project
+project_router = APIRouter(prefix="/project", tags=["project"])
+
+# Router for managing project resources
+project_resource_router = APIRouter(prefix="/projects", tags=["project_management"])
+
+
+@project_router.get("/info", response_model=ProjectInfoResponse)
+async def get_project_info(
+    project_service: ProjectServiceDep,
+    project: ProjectPathDep,
+) -> ProjectInfoResponse:
+    """Get comprehensive information about the specified Basic Memory project."""
+    return await project_service.get_project_info(project)
+
+
+# Update a project
+@project_router.patch("/{name}", response_model=ProjectStatusResponse)
+async def update_project(
+    project_service: ProjectServiceDep,
+    project_name: str = Path(..., description="Name of the project to update"),
+    path: Optional[str] = Body(None, description="New path for the project"),
+    is_active: Optional[bool] = Body(None, description="Status of the project (active/inactive)"),
+) -> ProjectStatusResponse:
+    """Update a project's information in configuration and database.
+
+    Args:
+        project_name: The name of the project to update
+        path: Optional new path for the project
+        is_active: Optional status update for the project
+
+    Returns:
+        Response confirming the project was updated
+    """
+    try:  # pragma: no cover
+        # Get original project info for the response
+        old_project_info = ProjectItem(
+            name=project_name,
+            path=project_service.projects.get(project_name, ""),
+        )
+
+        await project_service.update_project(project_name, updated_path=path, is_active=is_active)
+
+        # Get updated project info
+        updated_path = path if path else project_service.projects.get(project_name, "")
+
+        return ProjectStatusResponse(
+            message=f"Project '{project_name}' updated successfully",
+            status="success",
+            default=(project_name == project_service.default_project),
+            old_project=old_project_info,
+            new_project=ProjectItem(name=project_name, path=updated_path),
+        )
+    except ValueError as e:  # pragma: no cover
+        raise HTTPException(status_code=400, detail=str(e))
+
+
+# List all available projects
+@project_resource_router.get("/projects", response_model=ProjectList)
+async def list_projects(
+    project_service: ProjectServiceDep,
+) -> ProjectList:
+    """List all configured projects.
+
+    Returns:
+        A list of all projects with metadata
+    """
+    projects = await project_service.list_projects()
+    default_project = project_service.default_project
+
+    project_items = [
+        ProjectItem(
+            name=project.name,
+            path=project.path,
+            is_default=project.is_default or False,
+        )
+        for project in projects
+    ]
+
+    return ProjectList(
+        projects=project_items,
+        default_project=default_project,
+    )
+
+
+# Add a new project
+@project_resource_router.post("/projects", response_model=ProjectStatusResponse)
+async def add_project(
+    project_data: ProjectInfoRequest,
+    project_service: ProjectServiceDep,
+) -> ProjectStatusResponse:
+    """Add a new project to configuration and database.
+
+    Args:
+        project_data: The project name and path, with option to set as default
+
+    Returns:
+        Response confirming the project was added
+    """
+    try:  # pragma: no cover
+        await project_service.add_project(
+            project_data.name, project_data.path, set_default=project_data.set_default
+        )
+
+        return ProjectStatusResponse(  # pyright: ignore [reportCallIssue]
+            message=f"Project '{project_data.name}' added successfully",
+            status="success",
+            default=project_data.set_default,
+            new_project=ProjectItem(
+                name=project_data.name, path=project_data.path, is_default=project_data.set_default
+            ),
+        )
+    except ValueError as e:  # pragma: no cover
+        raise HTTPException(status_code=400, detail=str(e))
+
+
+# Remove a project
+@project_resource_router.delete("/{name}", response_model=ProjectStatusResponse)
+async def remove_project(
+    project_service: ProjectServiceDep,
+    name: str = Path(..., description="Name of the project to remove"),
+) -> ProjectStatusResponse:
+    """Remove a project from configuration and database.
+
+    Args:
+        name: The name of the project to remove
+
+    Returns:
+        Response confirming the project was removed
+    """
+    try:
+        old_project = await project_service.get_project(name)
+        if not old_project:  # pragma: no cover
+            raise HTTPException(
+                status_code=404, detail=f"Project: '{name}' does not exist"
+            )  # pragma: no cover
+
+        await project_service.remove_project(name)
+
+        return ProjectStatusResponse(
+            message=f"Project '{name}' removed successfully",
+            status="success",
+            default=False,
+            old_project=ProjectItem(name=old_project.name, path=old_project.path),
+            new_project=None,
+        )
+    except ValueError as e:  # pragma: no cover
+        raise HTTPException(status_code=400, detail=str(e))
+
+
+# Set a project as default
+@project_resource_router.put("/{name}/default", response_model=ProjectStatusResponse)
+async def set_default_project(
+    project_service: ProjectServiceDep,
+    name: str = Path(..., description="Name of the project to set as default"),
+) -> ProjectStatusResponse:
+    """Set a project as the default project.
+
+    Args:
+        name: The name of the project to set as default
+
+    Returns:
+        Response confirming the project was set as default
+    """
+    try:
+        # Get the old default project
+        default_name = project_service.default_project
+        default_project = await project_service.get_project(default_name)
+        if not default_project:  # pragma: no cover
+            raise HTTPException(  # pragma: no cover
+                status_code=404, detail=f"Default Project: '{default_name}' does not exist"
+            )
+
+        # get the new project
+        new_default_project = await project_service.get_project(name)
+        if not new_default_project:  # pragma: no cover
+            raise HTTPException(
+                status_code=404, detail=f"Project: '{name}' does not exist"
+            )  # pragma: no cover
+
+        await project_service.set_default_project(name)
+
+        return ProjectStatusResponse(
+            message=f"Project '{name}' set as default successfully",
+            status="success",
+            default=True,
+            old_project=ProjectItem(name=default_name, path=default_project.path),
+            new_project=ProjectItem(
+                name=name,
+                path=new_default_project.path,
+                is_default=True,
+            ),
+        )
+    except ValueError as e:  # pragma: no cover
+        raise HTTPException(status_code=400, detail=str(e))
+
+
+# Synchronize projects between config and database
+@project_resource_router.post("/sync", response_model=ProjectStatusResponse)
+async def synchronize_projects(
+    project_service: ProjectServiceDep,
+) -> ProjectStatusResponse:
+    """Synchronize projects between configuration file and database.
+
+    Ensures that all projects in the configuration file exist in the database
+    and vice versa.
+
+    Returns:
+        Response confirming synchronization was completed
+    """
+    try:  # pragma: no cover
+        await project_service.synchronize_projects()
+
+        return ProjectStatusResponse(  # pyright: ignore [reportCallIssue]
+            message="Projects synchronized successfully between configuration and database",
+            status="success",
+            default=False,
+        )
+    except ValueError as e:  # pragma: no cover
+        raise HTTPException(status_code=400, detail=str(e))

src/basic_memory/api/routers/prompt_router.py

@@ -1,22 +1,21 @@
-"""V2 Prompt Router - ID-based prompt generation operations.
+"""Router for prompt-related operations.
 
-This router uses v2 dependencies for consistent project handling with external_id UUIDs.
-Prompt endpoints are action-based (not resource-based), so they don't
-have entity IDs in URLs - they generate formatted prompts from queries.
+This router is responsible for rendering various prompts using Handlebars templates.
+It centralizes all prompt formatting logic that was previously in the MCP prompts.
 """
 
 from datetime import datetime, timezone
-from fastapi import APIRouter, HTTPException, status, Path
+from fastapi import APIRouter, HTTPException, status
 from loguru import logger
 
-from basic_memory.api.v2.utils import to_graph_context, to_search_results
+from basic_memory.api.routers.utils import to_graph_context, to_search_results
 from basic_memory.api.template_loader import template_loader
 from basic_memory.schemas.base import parse_timeframe
 from basic_memory.deps import (
-    ContextServiceV2ExternalDep,
-    EntityRepositoryV2ExternalDep,
-    SearchServiceV2ExternalDep,
-    EntityServiceV2ExternalDep,
+    ContextServiceDep,
+    EntityRepositoryDep,
+    SearchServiceDep,
+    EntityServiceDep,
 )
 from basic_memory.schemas.prompt import (
     ContinueConversationRequest,
@@ -26,17 +25,16 @@ from basic_memory.schemas.prompt import (
 )
 from basic_memory.schemas.search import SearchItemType, SearchQuery
 
-router = APIRouter(prefix="/prompt", tags=["prompt-v2"])
+router = APIRouter(prefix="/prompt", tags=["prompt"])
 
 
 @router.post("/continue-conversation", response_model=PromptResponse)
 async def continue_conversation(
-    search_service: SearchServiceV2ExternalDep,
-    entity_service: EntityServiceV2ExternalDep,
-    context_service: ContextServiceV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
+    search_service: SearchServiceDep,
+    entity_service: EntityServiceDep,
+    context_service: ContextServiceDep,
+    entity_repository: EntityRepositoryDep,
     request: ContinueConversationRequest,
-    project_id: str = Path(..., description="Project external UUID"),
 ) -> PromptResponse:
     """Generate a prompt for continuing a conversation.
 
@@ -44,15 +42,13 @@ async def continue_conversation(
     relevant context from the knowledge base.
 
     Args:
-        project_id: Project external UUID from URL path
         request: The request parameters
 
     Returns:
         Formatted continuation prompt with context
     """
     logger.info(
-        f"V2 Generating continue conversation prompt for project {project_id}, "
-        f"topic: {request.topic}, timeframe: {request.timeframe}"
+        f"Generating continue conversation prompt, topic: {request.topic}, timeframe: {request.timeframe}"
     )
 
     since = parse_timeframe(request.timeframe) if request.timeframe else None
@@ -196,10 +192,9 @@ async def continue_conversation(
 
 @router.post("/search", response_model=PromptResponse)
 async def search_prompt(
-    search_service: SearchServiceV2ExternalDep,
-    entity_service: EntityServiceV2ExternalDep,
+    search_service: SearchServiceDep,
+    entity_service: EntityServiceDep,
     request: SearchPromptRequest,
-    project_id: str = Path(..., description="Project external UUID"),
     page: int = 1,
     page_size: int = 10,
 ) -> PromptResponse:
@@ -209,7 +204,6 @@ async def search_prompt(
     prompt with context and suggestions.
 
     Args:
-        project_id: Project external UUID from URL path
         request: The search parameters
         page: The page number for pagination
         page_size: The number of results per page, defaults to 10
@@ -217,10 +211,7 @@ async def search_prompt(
     Returns:
         Formatted search results prompt with context
     """
-    logger.info(
-        f"V2 Generating search prompt for project {project_id}, "
-        f"query: {request.query}, timeframe: {request.timeframe}"
-    )
+    logger.info(f"Generating search prompt, query: {request.query}, timeframe: {request.timeframe}")
 
     limit = page_size
     offset = (page - 1) * page_size

src/basic_memory/api/routers/resource_router.py

@@ -0,0 +1,225 @@
+"""Routes for getting entity content."""
+
+import tempfile
+from pathlib import Path
+from typing import Annotated
+
+from fastapi import APIRouter, HTTPException, BackgroundTasks, Body
+from fastapi.responses import FileResponse, JSONResponse
+from loguru import logger
+
+from basic_memory.deps import (
+    ProjectConfigDep,
+    LinkResolverDep,
+    SearchServiceDep,
+    EntityServiceDep,
+    FileServiceDep,
+    EntityRepositoryDep,
+)
+from basic_memory.repository.search_repository import SearchIndexRow
+from basic_memory.schemas.memory import normalize_memory_url
+from basic_memory.schemas.search import SearchQuery, SearchItemType
+from basic_memory.models.knowledge import Entity as EntityModel
+from datetime import datetime
+
+router = APIRouter(prefix="/resource", tags=["resources"])
+
+
+def get_entity_ids(item: SearchIndexRow) -> set[int]:
+    match item.type:
+        case SearchItemType.ENTITY:
+            return {item.id}
+        case SearchItemType.OBSERVATION:
+            return {item.entity_id}  # pyright: ignore [reportReturnType]
+        case SearchItemType.RELATION:
+            from_entity = item.from_id
+            to_entity = item.to_id  # pyright: ignore [reportReturnType]
+            return {from_entity, to_entity} if to_entity else {from_entity}  # pyright: ignore [reportReturnType]
+        case _:  # pragma: no cover
+            raise ValueError(f"Unexpected type: {item.type}")
+
+
+@router.get("/{identifier:path}")
+async def get_resource_content(
+    config: ProjectConfigDep,
+    link_resolver: LinkResolverDep,
+    search_service: SearchServiceDep,
+    entity_service: EntityServiceDep,
+    file_service: FileServiceDep,
+    background_tasks: BackgroundTasks,
+    identifier: str,
+    page: int = 1,
+    page_size: int = 10,
+) -> FileResponse:
+    """Get resource content by identifier: name or permalink."""
+    logger.debug(f"Getting content for: {identifier}")
+
+    # Find single entity by permalink
+    entity = await link_resolver.resolve_link(identifier)
+    results = [entity] if entity else []
+
+    # pagination for multiple results
+    limit = page_size
+    offset = (page - 1) * page_size
+
+    # search using the identifier as a permalink
+    if not results:
+        # if the identifier contains a wildcard, use GLOB search
+        query = (
+            SearchQuery(permalink_match=identifier)
+            if "*" in identifier
+            else SearchQuery(permalink=identifier)
+        )
+        search_results = await search_service.search(query, limit, offset)
+        if not search_results:
+            raise HTTPException(status_code=404, detail=f"Resource not found: {identifier}")
+
+        # get the deduplicated entities related to the search results
+        entity_ids = {id for result in search_results for id in get_entity_ids(result)}
+        results = await entity_service.get_entities_by_id(list(entity_ids))
+
+    # return single response
+    if len(results) == 1:
+        entity = results[0]
+        file_path = Path(f"{config.home}/{entity.file_path}")
+        if not file_path.exists():
+            raise HTTPException(
+                status_code=404,
+                detail=f"File not found: {file_path}",
+            )
+        return FileResponse(path=file_path)
+
+    # for multiple files, initialize a temporary file for writing the results
+    with tempfile.NamedTemporaryFile(delete=False, mode="w", suffix=".md") as tmp_file:
+        temp_file_path = tmp_file.name
+
+        for result in results:
+            # Read content for each entity
+            content = await file_service.read_entity_content(result)
+            memory_url = normalize_memory_url(result.permalink)
+            modified_date = result.updated_at.isoformat()
+            checksum = result.checksum[:8] if result.checksum else ""
+
+            # Prepare the delimited content
+            response_content = f"--- {memory_url} {modified_date} {checksum}\n"
+            response_content += f"\n{content}\n"
+            response_content += "\n"
+
+            # Write content directly to the temporary file in append mode
+            tmp_file.write(response_content)
+
+        # Ensure all content is written to disk
+        tmp_file.flush()
+
+    # Schedule the temporary file to be deleted after the response
+    background_tasks.add_task(cleanup_temp_file, temp_file_path)
+
+    # Return the file response
+    return FileResponse(path=temp_file_path)
+
+
+def cleanup_temp_file(file_path: str):
+    """Delete the temporary file."""
+    try:
+        Path(file_path).unlink()  # Deletes the file
+        logger.debug(f"Temporary file deleted: {file_path}")
+    except Exception as e:  # pragma: no cover
+        logger.error(f"Error deleting temporary file {file_path}: {e}")
+
+
+@router.put("/{file_path:path}")
+async def write_resource(
+    config: ProjectConfigDep,
+    file_service: FileServiceDep,
+    entity_repository: EntityRepositoryDep,
+    search_service: SearchServiceDep,
+    file_path: str,
+    content: Annotated[str, Body()],
+) -> JSONResponse:
+    """Write content to a file in the project.
+
+    This endpoint allows writing content directly to a file in the project.
+    Also creates an entity record and indexes the file for search.
+
+    Args:
+        file_path: Path to write to, relative to project root
+        request: Contains the content to write
+
+    Returns:
+        JSON response with file information
+    """
+    try:
+        # Get content from request body
+
+        # Ensure it's UTF-8 string content
+        if isinstance(content, bytes):  # pragma: no cover
+            content_str = content.decode("utf-8")
+        else:
+            content_str = str(content)
+
+        # Get full file path
+        full_path = Path(f"{config.home}/{file_path}")
+
+        # Ensure parent directory exists
+        full_path.parent.mkdir(parents=True, exist_ok=True)
+
+        # Write content to file
+        checksum = await file_service.write_file(full_path, content_str)
+
+        # Get file info
+        file_stats = file_service.file_stats(full_path)
+
+        # Determine file details
+        file_name = Path(file_path).name
+        content_type = file_service.content_type(full_path)
+
+        entity_type = "canvas" if file_path.endswith(".canvas") else "file"
+
+        # Check if entity already exists
+        existing_entity = await entity_repository.get_by_file_path(file_path)
+
+        if existing_entity:
+            # Update existing entity
+            entity = await entity_repository.update(
+                existing_entity.id,
+                {
+                    "title": file_name,
+                    "entity_type": entity_type,
+                    "content_type": content_type,
+                    "file_path": file_path,
+                    "checksum": checksum,
+                    "updated_at": datetime.fromtimestamp(file_stats.st_mtime),
+                },
+            )
+            status_code = 200
+        else:
+            # Create a new entity model
+            entity = EntityModel(
+                title=file_name,
+                entity_type=entity_type,
+                content_type=content_type,
+                file_path=file_path,
+                checksum=checksum,
+                created_at=datetime.fromtimestamp(file_stats.st_ctime),
+                updated_at=datetime.fromtimestamp(file_stats.st_mtime),
+            )
+            entity = await entity_repository.add(entity)
+            status_code = 201
+
+        # Index the file for search
+        await search_service.index_entity(entity)  # pyright: ignore
+
+        # Return success response
+        return JSONResponse(
+            status_code=status_code,
+            content={
+                "file_path": file_path,
+                "checksum": checksum,
+                "size": file_stats.st_size,
+                "created_at": file_stats.st_ctime,
+                "modified_at": file_stats.st_mtime,
+            },
+        )
+    except Exception as e:  # pragma: no cover
+        logger.error(f"Error writing resource {file_path}: {e}")
+        raise HTTPException(status_code=500, detail=f"Failed to write resource: {str(e)}")

src/basic_memory/api/routers/search_router.py

@@ -0,0 +1,36 @@
+"""Router for search operations."""
+
+from fastapi import APIRouter, BackgroundTasks
+
+from basic_memory.api.routers.utils import to_search_results
+from basic_memory.schemas.search import SearchQuery, SearchResponse
+from basic_memory.deps import SearchServiceDep, EntityServiceDep
+
+router = APIRouter(prefix="/search", tags=["search"])
+
+
+@router.post("/", response_model=SearchResponse)
+async def search(
+    query: SearchQuery,
+    search_service: SearchServiceDep,
+    entity_service: EntityServiceDep,
+    page: int = 1,
+    page_size: int = 10,
+):
+    """Search across all knowledge and documents."""
+    limit = page_size
+    offset = (page - 1) * page_size
+    results = await search_service.search(query, limit=limit, offset=offset)
+    search_results = await to_search_results(entity_service, results)
+    return SearchResponse(
+        results=search_results,
+        current_page=page,
+        page_size=page_size,
+    )
+
+
+@router.post("/reindex")
+async def reindex(background_tasks: BackgroundTasks, search_service: SearchServiceDep):
+    """Recreate and populate the search index."""
+    await search_service.reindex_all(background_tasks=background_tasks)
+    return {"status": "ok", "message": "Reindex initiated"}

src/basic_memory/api/routers/utils.py

@@ -24,43 +24,11 @@ async def to_graph_context(
     page: Optional[int] = None,
     page_size: Optional[int] = None,
 ):
-    # First pass: collect all entity IDs needed for external_id lookup
-    # This includes: entity primary results, observation parent entities, relation from/to entities
-    entity_ids_needed: set[int] = set()
-    for context_item in context_result.results:
-        for item in (
-            [context_item.primary_result] + context_item.observations + context_item.related_results
-        ):
-            if item.type == SearchItemType.ENTITY:
-                # Entity's own ID for its external_id
-                entity_ids_needed.add(item.id)
-            elif item.type == SearchItemType.OBSERVATION:
-                # Parent entity ID for entity_external_id
-                if item.entity_id:  # pyright: ignore
-                    entity_ids_needed.add(item.entity_id)  # pyright: ignore
-            elif item.type == SearchItemType.RELATION:
-                # Source and target entity IDs for external_ids
-                if item.from_id:  # pyright: ignore
-                    entity_ids_needed.add(item.from_id)  # pyright: ignore
-                if item.to_id:
-                    entity_ids_needed.add(item.to_id)
-
-    # Batch fetch all entities at once - get both title and external_id
-    entity_title_lookup: dict[int, str] = {}
-    entity_external_id_lookup: dict[int, str] = {}
-    if entity_ids_needed:
-        entities = await entity_repository.find_by_ids(list(entity_ids_needed))
-        for e in entities:
-            entity_title_lookup[e.id] = e.title
-            entity_external_id_lookup[e.id] = e.external_id
-
     # Helper function to convert items to summaries
-    def to_summary(item: SearchIndexRow | ContextResultRow):
+    async def to_summary(item: SearchIndexRow | ContextResultRow):
         match item.type:
             case SearchItemType.ENTITY:
                 return EntitySummary(
-                    external_id=entity_external_id_lookup.get(item.id, ""),
-                    entity_id=item.id,
                     title=item.title,  # pyright: ignore
                     permalink=item.permalink,
                     content=item.content,
@@ -68,14 +36,8 @@ async def to_graph_context(
                     created_at=item.created_at,
                 )
             case SearchItemType.OBSERVATION:
-                entity_ext_id = None
-                if item.entity_id:  # pyright: ignore
-                    entity_ext_id = entity_external_id_lookup.get(item.entity_id)  # pyright: ignore
                 return ObservationSummary(
-                    observation_id=item.id,
-                    entity_id=item.entity_id,  # pyright: ignore
-                    entity_external_id=entity_ext_id,
-                    title=entity_title_lookup.get(item.entity_id),  # pyright: ignore
+                    title=item.title,  # pyright: ignore
                     file_path=item.file_path,
                     category=item.category,  # pyright: ignore
                     content=item.content,  # pyright: ignore
@@ -83,23 +45,15 @@ async def to_graph_context(
                     created_at=item.created_at,
                 )
             case SearchItemType.RELATION:
-                from_title = entity_title_lookup.get(item.from_id) if item.from_id else None  # pyright: ignore
-                to_title = entity_title_lookup.get(item.to_id) if item.to_id else None
-                from_ext_id = entity_external_id_lookup.get(item.from_id) if item.from_id else None  # pyright: ignore
-                to_ext_id = entity_external_id_lookup.get(item.to_id) if item.to_id else None
+                from_entity = await entity_repository.find_by_id(item.from_id)  # pyright: ignore
+                to_entity = await entity_repository.find_by_id(item.to_id) if item.to_id else None
                 return RelationSummary(
-                    relation_id=item.id,
-                    entity_id=item.entity_id,  # pyright: ignore
                     title=item.title,  # pyright: ignore
                     file_path=item.file_path,
                     permalink=item.permalink,  # pyright: ignore
                     relation_type=item.relation_type,  # pyright: ignore
-                    from_entity=from_title,
-                    from_entity_id=item.from_id,  # pyright: ignore
-                    from_entity_external_id=from_ext_id,
-                    to_entity=to_title,
-                    to_entity_id=item.to_id,
-                    to_entity_external_id=to_ext_id,
+                    from_entity=from_entity.title if from_entity else None,
+                    to_entity=to_entity.title if to_entity else None,
                     created_at=item.created_at,
                 )
             case _:  # pragma: no cover
@@ -109,19 +63,23 @@ async def to_graph_context(
     hierarchical_results = []
     for context_item in context_result.results:
         # Process primary result
-        primary_result = to_summary(context_item.primary_result)
+        primary_result = await to_summary(context_item.primary_result)
 
-        # Process observations (always ObservationSummary, validated by context_service)
-        observations = [to_summary(obs) for obs in context_item.observations]
+        # Process observations
+        observations = []
+        for obs in context_item.observations:
+            observations.append(await to_summary(obs))
 
         # Process related results
-        related = [to_summary(rel) for rel in context_item.related_results]
+        related = []
+        for rel in context_item.related_results:
+            related.append(await to_summary(rel))
 
         # Add to hierarchical results
         hierarchical_results.append(
             ContextResult(
                 primary_result=primary_result,
-                observations=observations,  # pyright: ignore[reportArgumentType]
+                observations=observations,
                 related_results=related,
             )
         )
@@ -146,7 +104,6 @@ async def to_graph_context(
         metadata=metadata,
         page=page,
         page_size=page_size,
-        has_more=context_result.metadata.has_more,
     )
 
 
@@ -154,21 +111,6 @@ async def to_search_results(entity_service: EntityService, results: List[SearchI
     search_results = []
     for r in results:
         entities = await entity_service.get_entities_by_id([r.entity_id, r.from_id, r.to_id])  # pyright: ignore
-
-        # Determine which IDs to set based on type
-        entity_id = None
-        observation_id = None
-        relation_id = None
-
-        if r.type == SearchItemType.ENTITY:
-            entity_id = r.id
-        elif r.type == SearchItemType.OBSERVATION:
-            observation_id = r.id
-            entity_id = r.entity_id  # Parent entity
-        elif r.type == SearchItemType.RELATION:
-            relation_id = r.id
-            entity_id = r.entity_id  # Parent entity
-
         search_results.append(
             SearchResult(
                 title=r.title,  # pyright: ignore
@@ -177,12 +119,8 @@ async def to_search_results(entity_service: EntityService, results: List[SearchI
                 score=r.score,  # pyright: ignore
                 entity=entities[0].permalink if entities else None,
                 content=r.content,
-                matched_chunk=r.matched_chunk_text,
                 file_path=r.file_path,
                 metadata=r.metadata,
-                entity_id=entity_id,
-                observation_id=observation_id,
-                relation_id=relation_id,
                 category=r.category,
                 from_entity=entities[0].permalink if entities else None,
                 to_entity=entities[1].permalink if len(entities) > 1 else None,

src/basic_memory/api/v2/__init__.py

@@ -1,35 +0,0 @@
-"""API v2 module - ID-based entity references.
-
-Version 2 of the Basic Memory API uses integer entity IDs as the primary
-identifier for improved performance and stability.
-
-Key changes from v1:
-- Entity lookups use integer IDs instead of paths/permalinks
-- Direct database queries instead of cascading resolution
-- Stable references that don't change with file moves
-- Better caching support
-
-All v2 routers are registered with the /v2 prefix.
-"""
-
-from basic_memory.api.v2.routers import (
-    knowledge_router,
-    memory_router,
-    project_router,
-    resource_router,
-    search_router,
-    directory_router,
-    prompt_router,
-    importer_router,
-)
-
-__all__ = [
-    "knowledge_router",
-    "memory_router",
-    "project_router",
-    "resource_router",
-    "search_router",
-    "directory_router",
-    "prompt_router",
-    "importer_router",
-]

src/basic_memory/api/v2/routers/__init__.py

@@ -1,23 +0,0 @@
-"""V2 API routers."""
-
-from basic_memory.api.v2.routers.knowledge_router import router as knowledge_router
-from basic_memory.api.v2.routers.project_router import router as project_router
-from basic_memory.api.v2.routers.memory_router import router as memory_router
-from basic_memory.api.v2.routers.search_router import router as search_router
-from basic_memory.api.v2.routers.resource_router import router as resource_router
-from basic_memory.api.v2.routers.directory_router import router as directory_router
-from basic_memory.api.v2.routers.prompt_router import router as prompt_router
-from basic_memory.api.v2.routers.importer_router import router as importer_router
-from basic_memory.api.v2.routers.schema_router import router as schema_router
-
-__all__ = [
-    "knowledge_router",
-    "project_router",
-    "memory_router",
-    "search_router",
-    "resource_router",
-    "directory_router",
-    "prompt_router",
-    "importer_router",
-    "schema_router",
-]

src/basic_memory/api/v2/routers/directory_router.py

@@ -1,93 +0,0 @@
-"""V2 Directory Router - ID-based directory tree operations.
-
-This router provides directory structure browsing for projects using
-external_id UUIDs instead of name-based identifiers.
-
-Key improvements:
-- Direct project lookup via external_id UUIDs
-- Consistent with other v2 endpoints
-- Better performance through indexed queries
-"""
-
-from typing import List, Optional
-
-from fastapi import APIRouter, Query, Path
-
-from basic_memory.deps import DirectoryServiceV2ExternalDep
-from basic_memory.schemas.directory import DirectoryNode
-
-router = APIRouter(prefix="/directory", tags=["directory-v2"])
-
-
-@router.get("/tree", response_model=DirectoryNode, response_model_exclude_none=True)
-async def get_directory_tree(
-    directory_service: DirectoryServiceV2ExternalDep,
-    project_id: str = Path(..., description="Project external UUID"),
-):
-    """Get hierarchical directory structure from the knowledge base.
-
-    Args:
-        directory_service: Service for directory operations
-        project_id: Project external UUID
-
-    Returns:
-        DirectoryNode representing the root of the hierarchical tree structure
-    """
-    # Get a hierarchical directory tree for the specific project
-    tree = await directory_service.get_directory_tree()
-
-    # Return the hierarchical tree
-    return tree
-
-
-@router.get("/structure", response_model=DirectoryNode, response_model_exclude_none=True)
-async def get_directory_structure(
-    directory_service: DirectoryServiceV2ExternalDep,
-    project_id: str = Path(..., description="Project external UUID"),
-):
-    """Get folder structure for navigation (no files).
-
-    Optimized endpoint for folder tree navigation. Returns only directory nodes
-    without file metadata. For full tree with files, use /directory/tree.
-
-    Args:
-        directory_service: Service for directory operations
-        project_id: Project external UUID
-
-    Returns:
-        DirectoryNode tree containing only folders (type="directory")
-    """
-    structure = await directory_service.get_directory_structure()
-    return structure
-
-
-@router.get("/list", response_model=List[DirectoryNode], response_model_exclude_none=True)
-async def list_directory(
-    directory_service: DirectoryServiceV2ExternalDep,
-    project_id: str = Path(..., description="Project external UUID"),
-    dir_name: str = Query("/", description="Directory path to list"),
-    depth: int = Query(1, ge=1, le=10, description="Recursion depth (1-10)"),
-    file_name_glob: Optional[str] = Query(
-        None, description="Glob pattern for filtering file names"
-    ),
-):
-    """List directory contents with filtering and depth control.
-
-    Args:
-        directory_service: Service for directory operations
-        project_id: Project external UUID
-        dir_name: Directory path to list (default: root "/")
-        depth: Recursion depth (1-10, default: 1 for immediate children only)
-        file_name_glob: Optional glob pattern for filtering file names (e.g., "*.md", "*meeting*")
-
-    Returns:
-        List of DirectoryNode objects matching the criteria
-    """
-    # Get directory listing with filtering
-    nodes = await directory_service.list_directory(
-        dir_name=dir_name,
-        depth=depth,
-        file_name_glob=file_name_glob,
-    )
-
-    return nodes

src/basic_memory/api/v2/routers/knowledge_router.py

@@ -1,684 +0,0 @@
-"""V2 Knowledge Router - External ID-based entity operations.
-
-This router provides external_id (UUID) based CRUD operations for entities,
-using stable string UUIDs that won't change with file moves or database migrations.
-
-Key improvements:
-- Stable external UUIDs that won't change with file moves or renames
-- Better API ergonomics with consistent string identifiers
-- Direct database lookups via unique indexed column
-- Simplified caching strategies
-"""
-
-from fastapi import APIRouter, HTTPException, BackgroundTasks, Depends, Response, Path, Query
-from loguru import logger
-
-from basic_memory.deps import (
-    EntityServiceV2ExternalDep,
-    SearchServiceV2ExternalDep,
-    LinkResolverV2ExternalDep,
-    ProjectConfigV2ExternalDep,
-    AppConfigDep,
-    EntityRepositoryV2ExternalDep,
-    RelationRepositoryV2ExternalDep,
-    ProjectExternalIdPathDep,
-    TaskSchedulerDep,
-    FileServiceV2ExternalDep,
-)
-from basic_memory.schemas import DeleteEntitiesResponse
-from basic_memory.schemas.base import Entity
-from basic_memory.schemas.request import EditEntityRequest
-from basic_memory.schemas.v2 import (
-    EntityResolveRequest,
-    EntityResolveResponse,
-    EntityResponseV2,
-    GraphEdge,
-    GraphNode,
-    GraphResponse,
-    MoveEntityRequestV2,
-    MoveDirectoryRequestV2,
-    DeleteDirectoryRequestV2,
-)
-from basic_memory.schemas.response import DirectoryMoveResult, DirectoryDeleteResult
-
-router = APIRouter(prefix="/knowledge", tags=["knowledge-v2"])
-
-
-def _schedule_vector_sync_if_enabled(
-    *,
-    task_scheduler,
-    app_config,
-    entity_id: int,
-    project_id: int,
-) -> None:
-    """Schedule out-of-band vector sync only when semantic search is enabled."""
-    if app_config.semantic_search_enabled:
-        task_scheduler.schedule(
-            "sync_entity_vectors",
-            entity_id=entity_id,
-            project_id=project_id,
-        )
-
-
-## Graph endpoint
-
-
-@router.get("/graph", response_model=GraphResponse)
-async def get_graph(
-    project_id: ProjectExternalIdPathDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    relation_repository: RelationRepositoryV2ExternalDep,
-) -> GraphResponse:
-    """Return all entities and resolved relations for knowledge graph visualization.
-
-    Returns a flat node/edge structure optimized for rendering with graph libraries.
-    Only includes resolved relations (where to_id is not null).
-    """
-    logger.info("API v2 request: get_graph")
-
-    # Fetch all entities for this project
-    entities = await entity_repository.find_all(use_load_options=False)
-    nodes = [
-        GraphNode(
-            external_id=entity.external_id,
-            title=entity.title,
-            note_type=entity.note_type,
-            file_path=entity.file_path,
-        )
-        for entity in entities
-    ]
-
-    # Fetch all resolved relations (to_id is not null) with eager-loaded entities
-    relations = await relation_repository.find_all()
-    edges = [
-        GraphEdge(
-            from_id=relation.from_entity.external_id,
-            to_id=relation.to_entity.external_id,
-            relation_type=relation.relation_type,
-        )
-        for relation in relations
-        if relation.to_entity is not None
-    ]
-
-    logger.info(f"API v2 response: graph with {len(nodes)} nodes and {len(edges)} edges")
-    return GraphResponse(nodes=nodes, edges=edges)
-
-
-## Resolution endpoint
-
-
-@router.post("/resolve", response_model=EntityResolveResponse)
-async def resolve_identifier(
-    project_id: ProjectExternalIdPathDep,
-    data: EntityResolveRequest,
-    link_resolver: LinkResolverV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-) -> EntityResolveResponse:
-    """Resolve a string identifier (external_id, permalink, title, or path) to entity info.
-
-    This endpoint provides a bridge between v1-style identifiers and v2 external_ids.
-    Use this to convert existing references to the new UUID-based format.
-
-    Args:
-        data: Request containing the identifier to resolve
-
-    Returns:
-        Entity external_id and metadata about how it was resolved
-
-    Raises:
-        HTTPException: 404 if identifier cannot be resolved
-
-    Example:
-        POST /v2/{project_id}/knowledge/resolve
-        {"identifier": "specs/search"}
-
-        Returns:
-        {
-            "external_id": "550e8400-e29b-41d4-a716-446655440000",
-            "entity_id": 123,
-            "permalink": "specs/search",
-            "file_path": "specs/search.md",
-            "title": "Search Specification",
-            "resolution_method": "permalink"
-        }
-    """
-    logger.info(f"API v2 request: resolve_identifier for '{data.identifier}'")
-
-    # Try to resolve by external_id first
-    entity = await entity_repository.get_by_external_id(data.identifier)
-    resolution_method = "external_id" if entity else "search"
-
-    # If not found by external_id, try other resolution methods
-    # Pass source_path for context-aware resolution (prefers notes closer to source)
-    # Pass strict to control fuzzy search fallback (default False allows fuzzy matching)
-    if not entity:
-        entity = await link_resolver.resolve_link(
-            data.identifier, source_path=data.source_path, strict=data.strict
-        )
-        if entity:
-            # Determine resolution method
-            if entity.permalink == data.identifier:
-                resolution_method = "permalink"
-            elif entity.title == data.identifier:
-                resolution_method = "title"
-            elif entity.file_path == data.identifier:
-                resolution_method = "path"
-            else:
-                resolution_method = "search"
-
-    if not entity:
-        raise HTTPException(status_code=404, detail=f"Entity not found: '{data.identifier}'")
-
-    result = EntityResolveResponse(
-        external_id=entity.external_id,
-        entity_id=entity.id,
-        permalink=entity.permalink,
-        file_path=entity.file_path,
-        title=entity.title,
-        resolution_method=resolution_method,
-    )
-
-    logger.debug(
-        f"API v2 response: resolved '{data.identifier}' to external_id={result.external_id} via {resolution_method}"
-    )
-
-    return result
-
-
-## Read endpoints
-
-
-@router.get("/entities/{entity_id}", response_model=EntityResponseV2)
-async def get_entity_by_id(
-    project_id: ProjectExternalIdPathDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    entity_id: str = Path(..., description="Entity external ID (UUID)"),
-) -> EntityResponseV2:
-    """Get an entity by its external ID (UUID).
-
-    This is the primary entity retrieval method in v2, using stable UUID
-    identifiers that won't change with file moves.
-
-    Args:
-        entity_id: External ID (UUID string)
-
-    Returns:
-        Complete entity with observations and relations
-
-    Raises:
-        HTTPException: 404 if entity not found
-    """
-    logger.info(f"API v2 request: get_entity_by_id entity_id={entity_id}")
-
-    entity = await entity_repository.get_by_external_id(entity_id)
-    if not entity:
-        raise HTTPException(
-            status_code=404, detail=f"Entity with external_id '{entity_id}' not found"
-        )
-
-    result = EntityResponseV2.model_validate(entity)
-    logger.info(f"API v2 response: external_id={entity_id}, title='{result.title}'")
-
-    return result
-
-
-## Create endpoints
-
-
-@router.post("/entities", response_model=EntityResponseV2)
-async def create_entity(
-    project_id: ProjectExternalIdPathDep,
-    data: Entity,
-    background_tasks: BackgroundTasks,
-    entity_service: EntityServiceV2ExternalDep,
-    search_service: SearchServiceV2ExternalDep,
-    task_scheduler: TaskSchedulerDep,
-    file_service: FileServiceV2ExternalDep,
-    app_config: AppConfigDep,
-    fast: bool = Query(
-        True, description="If true, write quickly and defer indexing to background tasks."
-    ),
-) -> EntityResponseV2:
-    """Create a new entity.
-
-    Args:
-        data: Entity data to create
-        fast: If True, defer indexing to background tasks
-
-    Returns:
-        Created entity with generated external_id (UUID) and file content
-    """
-    logger.info(
-        "API v2 request", endpoint="create_entity", note_type=data.note_type, title=data.title
-    )
-
-    if fast:
-        entity = await entity_service.fast_write_entity(data)
-        task_scheduler.schedule(
-            "reindex_entity",
-            entity_id=entity.id,
-            project_id=project_id,
-        )
-    else:
-        entity = await entity_service.create_entity(data)
-        await search_service.index_entity(entity)
-        _schedule_vector_sync_if_enabled(
-            task_scheduler=task_scheduler,
-            app_config=app_config,
-            entity_id=entity.id,
-            project_id=project_id,
-        )
-
-    result = EntityResponseV2.model_validate(entity)
-    if fast:
-        result = result.model_copy(update={"observations": [], "relations": []})
-
-    # Always read and return file content
-    content = await file_service.read_file_content(entity.file_path)
-    result = result.model_copy(update={"content": content})
-
-    logger.info(
-        f"API v2 response: endpoint='create_entity' external_id={entity.external_id}, title={result.title}, permalink={result.permalink}, status_code=201"
-    )
-    return result
-
-
-## Update endpoints
-
-
-@router.put("/entities/{entity_id}", response_model=EntityResponseV2)
-async def update_entity_by_id(
-    data: Entity,
-    response: Response,
-    background_tasks: BackgroundTasks,
-    project_id: ProjectExternalIdPathDep,
-    entity_service: EntityServiceV2ExternalDep,
-    search_service: SearchServiceV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    task_scheduler: TaskSchedulerDep,
-    file_service: FileServiceV2ExternalDep,
-    app_config: AppConfigDep,
-    entity_id: str = Path(..., description="Entity external ID (UUID)"),
-    fast: bool = Query(
-        True, description="If true, write quickly and defer indexing to background tasks."
-    ),
-) -> EntityResponseV2:
-    """Update an entity by external ID.
-
-    If the entity doesn't exist, it will be created (upsert behavior).
-
-    Args:
-        entity_id: External ID (UUID string)
-        data: Updated entity data
-        fast: If True, defer indexing to background tasks
-
-    Returns:
-        Updated entity with file content
-    """
-    logger.info(f"API v2 request: update_entity_by_id entity_id={entity_id}")
-
-    # Check if entity exists (external_id is the source of truth for v2)
-    existing = await entity_repository.get_by_external_id(entity_id)
-    created = existing is None
-
-    if fast:
-        entity = await entity_service.fast_write_entity(data, external_id=entity_id)
-        response.status_code = 200 if existing else 201
-        task_scheduler.schedule(
-            "reindex_entity",
-            entity_id=entity.id,
-            project_id=project_id,
-            resolve_relations=created,
-        )
-    else:
-        if existing:
-            # Update the existing entity in-place to avoid path-based duplication
-            entity = await entity_service.update_entity(existing, data)
-            response.status_code = 200
-        else:
-            # Create new entity, then bind external_id to the requested UUID
-            entity = await entity_service.create_entity(data)
-            if entity.external_id != entity_id:
-                entity = await entity_repository.update(
-                    entity.id,
-                    {"external_id": entity_id},
-                )
-                if not entity:
-                    raise HTTPException(
-                        status_code=404,
-                        detail=f"Entity with external_id '{entity_id}' not found",
-                    )
-            response.status_code = 201
-
-        await search_service.index_entity(entity)
-        _schedule_vector_sync_if_enabled(
-            task_scheduler=task_scheduler,
-            app_config=app_config,
-            entity_id=entity.id,
-            project_id=project_id,
-        )
-
-    result = EntityResponseV2.model_validate(entity)
-    if fast:
-        result = result.model_copy(update={"observations": [], "relations": []})
-
-    # Always read and return file content
-    content = await file_service.read_file_content(entity.file_path)
-    result = result.model_copy(update={"content": content})
-
-    logger.info(
-        f"API v2 response: external_id={entity_id}, created={created}, status_code={response.status_code}"
-    )
-    return result
-
-
-@router.patch("/entities/{entity_id}", response_model=EntityResponseV2)
-async def edit_entity_by_id(
-    data: EditEntityRequest,
-    background_tasks: BackgroundTasks,
-    project_id: ProjectExternalIdPathDep,
-    entity_service: EntityServiceV2ExternalDep,
-    search_service: SearchServiceV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    task_scheduler: TaskSchedulerDep,
-    file_service: FileServiceV2ExternalDep,
-    app_config: AppConfigDep,
-    entity_id: str = Path(..., description="Entity external ID (UUID)"),
-    fast: bool = Query(
-        True, description="If true, write quickly and defer indexing to background tasks."
-    ),
-) -> EntityResponseV2:
-    """Edit an existing entity by external ID using operations like append, prepend, etc.
-
-    Args:
-        entity_id: External ID (UUID string)
-        data: Edit operation details
-        fast: If True, defer indexing to background tasks
-
-    Returns:
-        Updated entity with file content
-
-    Raises:
-        HTTPException: 404 if entity not found, 400 if edit fails
-    """
-    logger.info(
-        f"API v2 request: edit_entity_by_id entity_id={entity_id}, operation='{data.operation}'"
-    )
-
-    # Verify entity exists
-    entity = await entity_repository.get_by_external_id(entity_id)
-    if not entity:  # pragma: no cover
-        raise HTTPException(
-            status_code=404, detail=f"Entity with external_id '{entity_id}' not found"
-        )
-
-    try:
-        if fast:
-            updated_entity = await entity_service.fast_edit_entity(
-                entity=entity,
-                operation=data.operation,
-                content=data.content,
-                section=data.section,
-                find_text=data.find_text,
-                expected_replacements=data.expected_replacements,
-            )
-            task_scheduler.schedule(
-                "reindex_entity",
-                entity_id=updated_entity.id,
-                project_id=project_id,
-            )
-        else:
-            # Edit using the entity's permalink or path
-            identifier = entity.permalink or entity.file_path
-            updated_entity = await entity_service.edit_entity(
-                identifier=identifier,
-                operation=data.operation,
-                content=data.content,
-                section=data.section,
-                find_text=data.find_text,
-                expected_replacements=data.expected_replacements,
-            )
-
-            await search_service.index_entity(updated_entity)
-            _schedule_vector_sync_if_enabled(
-                task_scheduler=task_scheduler,
-                app_config=app_config,
-                entity_id=updated_entity.id,
-                project_id=project_id,
-            )
-
-        result = EntityResponseV2.model_validate(updated_entity)
-        if fast:
-            result = result.model_copy(update={"observations": [], "relations": []})
-
-        # Always read and return file content
-        content = await file_service.read_file_content(updated_entity.file_path)
-        result = result.model_copy(update={"content": content})
-
-        logger.info(
-            f"API v2 response: external_id={entity_id}, operation='{data.operation}', status_code=200"
-        )
-
-        return result
-
-    except Exception as e:
-        logger.error(f"Error editing entity {entity_id}: {e}")
-        raise HTTPException(status_code=400, detail=str(e))
-
-
-## Delete endpoints
-
-
-@router.delete("/entities/{entity_id}", response_model=DeleteEntitiesResponse)
-async def delete_entity_by_id(
-    background_tasks: BackgroundTasks,
-    project_id: ProjectExternalIdPathDep,
-    entity_service: EntityServiceV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    entity_id: str = Path(..., description="Entity external ID (UUID)"),
-    search_service=Depends(lambda: None),  # Optional for now
-) -> DeleteEntitiesResponse:
-    """Delete an entity by external ID.
-
-    Args:
-        entity_id: External ID (UUID string)
-
-    Returns:
-        Deletion status
-
-    Note: Returns deleted=False if entity doesn't exist (idempotent)
-    """
-    logger.info(f"API v2 request: delete_entity_by_id entity_id={entity_id}")
-
-    entity = await entity_repository.get_by_external_id(entity_id)
-    if entity is None:
-        logger.info(f"API v2 response: external_id={entity_id} not found, deleted=False")
-        return DeleteEntitiesResponse(deleted=False)
-
-    # Delete the entity using internal ID
-    deleted = await entity_service.delete_entity(entity.id)
-
-    # Remove from search index if search service available
-    if search_service:
-        background_tasks.add_task(search_service.handle_delete, entity)  # pragma: no cover
-
-    logger.info(f"API v2 response: external_id={entity_id}, deleted={deleted}")
-
-    return DeleteEntitiesResponse(deleted=deleted)
-
-
-## Move endpoint
-
-
-@router.put("/entities/{entity_id}/move", response_model=EntityResponseV2)
-async def move_entity(
-    data: MoveEntityRequestV2,
-    background_tasks: BackgroundTasks,
-    project_id: ProjectExternalIdPathDep,
-    entity_service: EntityServiceV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    project_config: ProjectConfigV2ExternalDep,
-    app_config: AppConfigDep,
-    search_service: SearchServiceV2ExternalDep,
-    task_scheduler: TaskSchedulerDep,
-    entity_id: str = Path(..., description="Entity external ID (UUID)"),
-) -> EntityResponseV2:
-    """Move an entity to a new file location.
-
-    V2 API uses external_id (UUID) in the URL path for stable references.
-    The external_id will remain stable after the move.
-
-    Args:
-        project_id: Project external ID from URL path
-        entity_id: Entity external ID from URL path (primary identifier)
-        data: Move request with destination path only
-
-    Returns:
-        Updated entity with new file path
-    """
-    logger.info(
-        f"API v2 request: move_entity entity_id={entity_id}, destination='{data.destination_path}'"
-    )
-
-    try:
-        # First, get the entity by external_id to verify it exists
-        entity = await entity_repository.get_by_external_id(entity_id)
-        if not entity:  # pragma: no cover
-            raise HTTPException(
-                status_code=404, detail=f"Entity with external_id '{entity_id}' not found"
-            )
-
-        # Move the entity using its current file path as identifier
-        moved_entity = await entity_service.move_entity(
-            identifier=entity.file_path,  # Use file path for resolution
-            destination_path=data.destination_path,
-            project_config=project_config,
-            app_config=app_config,
-        )
-
-        # Reindex at new location
-        reindexed_entity = await entity_service.link_resolver.resolve_link(data.destination_path)
-        if reindexed_entity:
-            await search_service.index_entity(reindexed_entity)
-            _schedule_vector_sync_if_enabled(
-                task_scheduler=task_scheduler,
-                app_config=app_config,
-                entity_id=reindexed_entity.id,
-                project_id=project_id,
-            )
-
-        result = EntityResponseV2.model_validate(moved_entity)
-
-        logger.info(f"API v2 response: moved external_id={entity_id} to '{data.destination_path}'")
-
-        return result
-
-    except HTTPException:  # pragma: no cover
-        raise  # pragma: no cover
-    except Exception as e:
-        logger.error(f"Error moving entity: {e}")
-        raise HTTPException(status_code=400, detail=str(e))
-
-
-## Move directory endpoint
-
-
-@router.post("/move-directory", response_model=DirectoryMoveResult)
-async def move_directory(
-    data: MoveDirectoryRequestV2,
-    background_tasks: BackgroundTasks,
-    project_id: ProjectExternalIdPathDep,
-    entity_service: EntityServiceV2ExternalDep,
-    project_config: ProjectConfigV2ExternalDep,
-    app_config: AppConfigDep,
-    search_service: SearchServiceV2ExternalDep,
-    task_scheduler: TaskSchedulerDep,
-) -> DirectoryMoveResult:
-    """Move all entities in a directory to a new location.
-
-    V2 API uses project external_id in the URL path for stable references.
-    Moves all files within a source directory to a destination directory,
-    updating database records and optionally updating permalinks.
-
-    Args:
-        project_id: Project external ID from URL path
-        data: Move request with source and destination directories
-
-    Returns:
-        DirectoryMoveResult with counts and details of moved files
-    """
-    logger.info(
-        f"API v2 request: move_directory source='{data.source_directory}', destination='{data.destination_directory}'"
-    )
-
-    try:
-        # Move the directory using the service
-        result = await entity_service.move_directory(
-            source_directory=data.source_directory,
-            destination_directory=data.destination_directory,
-            project_config=project_config,
-            app_config=app_config,
-        )
-
-        # Reindex moved entities
-        for file_path in result.moved_files:
-            entity = await entity_service.link_resolver.resolve_link(file_path)
-            if entity:
-                await search_service.index_entity(entity)
-                _schedule_vector_sync_if_enabled(
-                    task_scheduler=task_scheduler,
-                    app_config=app_config,
-                    entity_id=entity.id,
-                    project_id=project_id,
-                )
-
-        logger.info(
-            f"API v2 response: move_directory "
-            f"total={result.total_files}, success={result.successful_moves}, failed={result.failed_moves}"
-        )
-        return result
-
-    except Exception as e:
-        logger.error(f"Error moving directory: {e}")
-        raise HTTPException(status_code=400, detail=str(e))
-
-
-## Delete directory endpoint
-
-
-@router.post("/delete-directory", response_model=DirectoryDeleteResult)
-async def delete_directory(
-    data: DeleteDirectoryRequestV2,
-    project_id: ProjectExternalIdPathDep,
-    entity_service: EntityServiceV2ExternalDep,
-) -> DirectoryDeleteResult:
-    """Delete all entities in a directory.
-
-    V2 API uses project external_id in the URL path for stable references.
-    Deletes all files within a directory, updating database records and
-    removing files from the filesystem.
-
-    Args:
-        project_id: Project external ID from URL path
-        data: Delete request with directory path
-
-    Returns:
-        DirectoryDeleteResult with counts and details of deleted files
-    """
-    logger.info(f"API v2 request: delete_directory directory='{data.directory}'")
-
-    try:
-        # Delete the directory using the service
-        result = await entity_service.delete_directory(
-            directory=data.directory,
-        )
-
-        logger.info(
-            f"API v2 response: delete_directory "
-            f"total={result.total_files}, success={result.successful_deletes}, failed={result.failed_deletes}"
-        )
-        return result
-
-    except Exception as e:
-        logger.error(f"Error deleting directory: {e}")
-        raise HTTPException(status_code=400, detail=str(e))

src/basic_memory/api/v2/routers/memory_router.py

@@ -1,130 +0,0 @@
-"""V2 routes for memory:// URI operations.
-
-This router uses external_id UUIDs for stable, API-friendly routing.
-V1 uses string-based project names which are less efficient and less stable.
-"""
-
-from typing import Annotated, Optional
-
-from fastapi import APIRouter, Query, Path
-from loguru import logger
-
-from basic_memory.deps import ContextServiceV2ExternalDep, EntityRepositoryV2ExternalDep
-from basic_memory.schemas.base import TimeFrame, parse_timeframe
-from basic_memory.schemas.memory import (
-    GraphContext,
-    normalize_memory_url,
-)
-from basic_memory.schemas.search import SearchItemType
-from basic_memory.api.v2.utils import to_graph_context
-
-# Note: No prefix here - it's added during registration as /v2/{project_id}/memory
-router = APIRouter(tags=["memory"])
-
-
-@router.get("/memory/recent", response_model=GraphContext)
-async def recent(
-    context_service: ContextServiceV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    project_id: str = Path(..., description="Project external UUID"),
-    type: Annotated[list[SearchItemType] | None, Query()] = None,
-    depth: int = 1,
-    timeframe: TimeFrame = "7d",
-    page: int = 1,
-    page_size: int = 10,
-    max_related: int = 10,
-) -> GraphContext:
-    """Get recent activity context for a project.
-
-    Args:
-        project_id: Project external UUID from URL path
-        context_service: Context service scoped to project
-        entity_repository: Entity repository scoped to project
-        type: Types of items to include (entities, relations, observations)
-        depth: How many levels of related entities to include
-        timeframe: Time window for recent activity (e.g., "7d", "1 week")
-        page: Page number for pagination
-        page_size: Number of items per page
-        max_related: Maximum related entities to include per item
-
-    Returns:
-        GraphContext with recent activity and related entities
-    """
-    # return all types by default
-    types = (
-        [SearchItemType.ENTITY, SearchItemType.RELATION, SearchItemType.OBSERVATION]
-        if not type
-        else type
-    )
-
-    logger.debug(
-        f"V2 Getting recent context for project {project_id}: `{types}` depth: `{depth}` timeframe: `{timeframe}` page: `{page}` page_size: `{page_size}` max_related: `{max_related}`"
-    )
-    # Parse timeframe
-    since = parse_timeframe(timeframe)
-    limit = page_size
-    offset = (page - 1) * page_size
-
-    # Build context
-    context = await context_service.build_context(
-        types=types, depth=depth, since=since, limit=limit, offset=offset, max_related=max_related
-    )
-    recent_context = await to_graph_context(
-        context, entity_repository=entity_repository, page=page, page_size=page_size
-    )
-    logger.debug(f"V2 Recent context: {recent_context.model_dump_json()}")
-    return recent_context
-
-
-# get_memory_context needs to be declared last so other paths can match
-
-
-@router.get("/memory/{uri:path}", response_model=GraphContext)
-async def get_memory_context(
-    context_service: ContextServiceV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    uri: str,
-    project_id: str = Path(..., description="Project external UUID"),
-    depth: int = 1,
-    timeframe: Optional[TimeFrame] = None,
-    page: int = 1,
-    page_size: int = 10,
-    max_related: int = 10,
-) -> GraphContext:
-    """Get rich context from memory:// URI.
-
-    V2 supports both legacy path-based URIs and new ID-based URIs:
-    - Legacy: memory://path/to/note
-    - ID-based: memory://id/123 or memory://123
-
-    Args:
-        project_id: Project external UUID from URL path
-        context_service: Context service scoped to project
-        entity_repository: Entity repository scoped to project
-        uri: Memory URI path (e.g., "id/123", "123", or "path/to/note")
-        depth: How many levels of related entities to include
-        timeframe: Optional time window for filtering related content
-        page: Page number for pagination
-        page_size: Number of items per page
-        max_related: Maximum related entities to include
-
-    Returns:
-        GraphContext with the entity and its related context
-    """
-    logger.debug(
-        f"V2 Getting context for project {project_id}, URI: `{uri}` depth: `{depth}` timeframe: `{timeframe}` page: `{page}` page_size: `{page_size}` max_related: `{max_related}`"
-    )
-    memory_url = normalize_memory_url(uri)
-
-    # Parse timeframe
-    since = parse_timeframe(timeframe) if timeframe else None
-    limit = page_size
-    offset = (page - 1) * page_size
-
-    # Build context
-    context = await context_service.build_context(
-        memory_url, depth=depth, since=since, limit=limit, offset=offset, max_related=max_related
-    )
-    return await to_graph_context(
-        context, entity_repository=entity_repository, page=page, page_size=page_size
-    )

src/basic_memory/api/v2/routers/project_router.py

@@ -1,550 +0,0 @@
-"""V2 Project Router - External ID-based project management operations.
-
-This router provides external_id (UUID) based CRUD operations for projects,
-using stable string UUIDs that never change (unlike integer IDs or names).
-
-Key improvements:
-- Stable external UUIDs that won't change with renames or database migrations
-- Better API ergonomics with consistent string identifiers
-- Direct database lookups via unique indexed column
-- Consistent with v2 entity operations
-"""
-
-import os
-from typing import Optional
-
-from fastapi import APIRouter, HTTPException, Body, Query, Path
-from loguru import logger
-
-from basic_memory.deps import (
-    ProjectServiceDep,
-    ProjectRepositoryDep,
-    ProjectConfigV2ExternalDep,
-    SyncServiceV2ExternalDep,
-    TaskSchedulerDep,
-    ProjectExternalIdPathDep,
-)
-from basic_memory.schemas import SyncReportResponse
-from basic_memory.schemas.project_info import (
-    ProjectItem,
-    ProjectList,
-    ProjectInfoRequest,
-    ProjectInfoResponse,
-    ProjectStatusResponse,
-)
-from basic_memory.schemas.v2 import ProjectResolveRequest, ProjectResolveResponse
-from basic_memory.utils import normalize_project_path, generate_permalink
-
-router = APIRouter(prefix="/projects", tags=["project_management-v2"])
-
-
-@router.get("/", response_model=ProjectList)
-async def list_projects(
-    project_service: ProjectServiceDep,
-) -> ProjectList:
-    """List all configured projects.
-
-    Returns:
-        A list of all projects with metadata
-    """
-    projects = await project_service.list_projects()
-    default_project = await project_service.get_default_project_name()
-
-    project_items = [
-        ProjectItem(
-            id=project.id,
-            external_id=project.external_id,
-            name=project.name,
-            path=normalize_project_path(project.path),
-            is_default=project.is_default or False,
-        )
-        for project in projects
-    ]
-
-    return ProjectList(
-        projects=project_items,
-        default_project=default_project,
-    )
-
-
-@router.post("/", response_model=ProjectStatusResponse, status_code=201)
-async def add_project(
-    project_data: ProjectInfoRequest,
-    project_service: ProjectServiceDep,
-) -> ProjectStatusResponse:
-    """Add a new project to configuration and database.
-
-    Args:
-        project_data: The project name and path, with option to set as default
-
-    Returns:
-        Response confirming the project was added
-    """
-    # Check if project already exists before attempting to add
-    existing_project = await project_service.get_project(project_data.name)
-    if existing_project:
-        # Project exists - check if paths match for true idempotency
-        # Normalize paths for comparison (resolve symlinks, etc.)
-        requested_path = os.path.abspath(os.path.expanduser(project_data.path))
-        existing_path = os.path.abspath(os.path.expanduser(existing_project.path))
-
-        if requested_path == existing_path:
-            # Same name, same path - return 200 OK (idempotent)
-            return ProjectStatusResponse(  # pyright: ignore [reportCallIssue]
-                message=f"Project '{project_data.name}' already exists",
-                status="success",
-                default=existing_project.is_default or False,
-                new_project=ProjectItem(
-                    id=existing_project.id,
-                    external_id=existing_project.external_id,
-                    name=existing_project.name,
-                    path=existing_project.path,
-                    is_default=existing_project.is_default or False,
-                ),
-            )
-        else:
-            # Same name, different path - this is an error
-            raise HTTPException(
-                status_code=400,
-                detail=(
-                    f"Project '{project_data.name}' already exists with different path. "
-                    f"Existing: {existing_project.path}, Requested: {project_data.path}"
-                ),
-            )
-
-    try:  # pragma: no cover
-        # The service layer handles cloud mode validation and path sanitization
-        await project_service.add_project(
-            project_data.name, project_data.path, set_default=project_data.set_default
-        )
-
-        # Fetch the newly created project to get its ID
-        new_project = await project_service.get_project(project_data.name)
-        if not new_project:
-            raise HTTPException(status_code=500, detail="Failed to retrieve newly created project")
-
-        return ProjectStatusResponse(  # pyright: ignore [reportCallIssue]
-            message=f"Project '{new_project.name}' added successfully",
-            status="success",
-            default=project_data.set_default,
-            new_project=ProjectItem(
-                id=new_project.id,
-                external_id=new_project.external_id,
-                name=new_project.name,
-                path=new_project.path,
-                is_default=new_project.is_default or False,
-            ),
-        )
-    except ValueError as e:  # pragma: no cover
-        raise HTTPException(status_code=400, detail=str(e))
-
-
-@router.post("/config/sync", response_model=ProjectStatusResponse)
-async def synchronize_projects(
-    project_service: ProjectServiceDep,
-) -> ProjectStatusResponse:
-    """Synchronize projects between configuration file and database."""
-    try:  # pragma: no cover
-        await project_service.synchronize_projects()
-
-        return ProjectStatusResponse(  # pyright: ignore [reportCallIssue]
-            message="Projects synchronized successfully between configuration and database",
-            status="success",
-            default=False,
-        )
-    except ValueError as e:  # pragma: no cover
-        raise HTTPException(status_code=400, detail=str(e))
-
-
-@router.post("/{project_id}/sync")
-async def sync_project(
-    sync_service: SyncServiceV2ExternalDep,
-    project_config: ProjectConfigV2ExternalDep,
-    task_scheduler: TaskSchedulerDep,
-    project_internal_id: ProjectExternalIdPathDep,
-    force_full: bool = Query(
-        False, description="Force full scan, bypassing watermark optimization"
-    ),
-    run_in_background: bool = Query(True, description="Run in background"),
-):
-    """Force project filesystem sync to database."""
-    if run_in_background:
-        task_scheduler.schedule(
-            "sync_project",
-            project_id=project_internal_id,
-            force_full=force_full,
-        )
-        logger.info(
-            f"Filesystem sync initiated for project: {project_config.name} (force_full={force_full})"
-        )
-
-        return {
-            "status": "sync_started",
-            "message": f"Filesystem sync initiated for project '{project_config.name}'",
-        }
-
-    report = await sync_service.sync(
-        project_config.home, project_config.name, force_full=force_full
-    )
-    logger.info(
-        f"Filesystem sync completed for project: {project_config.name} (force_full={force_full})"
-    )
-    return SyncReportResponse.from_sync_report(report)
-
-
-@router.post("/{project_id}/status", response_model=SyncReportResponse)
-async def get_project_status(
-    sync_service: SyncServiceV2ExternalDep,
-    project_config: ProjectConfigV2ExternalDep,
-    project_id: str = Path(..., description="Project external ID (UUID)"),
-    force_full: bool = Query(
-        False, description="Force full scan, bypassing watermark optimization"
-    ),
-) -> SyncReportResponse:
-    """Get sync status of files vs database for a project."""
-    logger.info(f"API v2 request: get_project_status for project_id={project_id}")
-    report = await sync_service.scan(project_config.home, force_full=force_full)
-    return SyncReportResponse.from_sync_report(report)
-
-
-@router.post("/resolve", response_model=ProjectResolveResponse)
-async def resolve_project_identifier(
-    data: ProjectResolveRequest,
-    project_repository: ProjectRepositoryDep,
-) -> ProjectResolveResponse:
-    """Resolve a project identifier (name, permalink, or external_id) to project info.
-
-    This endpoint provides efficient lookup of projects by various identifiers
-    without needing to fetch the entire project list. Supports:
-    - External ID (UUID string) - preferred stable identifier
-    - Permalink
-    - Case-insensitive name matching
-
-    Args:
-        data: Request containing the identifier to resolve
-
-    Returns:
-        Project information including the external_id (UUID)
-
-    Raises:
-        HTTPException: 404 if project not found
-
-    Example:
-        POST /v2/projects/resolve
-        {"identifier": "my-project"}
-
-        Returns:
-        {
-            "external_id": "550e8400-e29b-41d4-a716-446655440000",
-            "project_id": 1,
-            "name": "my-project",
-            "permalink": "my-project",
-            "path": "/path/to/project",
-            "is_active": true,
-            "is_default": false,
-            "resolution_method": "name"
-        }
-    """
-    logger.info(f"API v2 request: resolve_project_identifier for '{data.identifier}'")
-
-    # Generate permalink for comparison
-    identifier_permalink = generate_permalink(data.identifier)
-
-    resolution_method = "name"
-    project = None
-
-    # Try external_id first (UUID format)
-    project = await project_repository.get_by_external_id(data.identifier)
-    if project:
-        resolution_method = "external_id"
-
-    # If not found by external_id, try by permalink (exact match)
-    if not project:
-        project = await project_repository.get_by_permalink(identifier_permalink)
-        if project:
-            resolution_method = "permalink"
-
-    # If not found by permalink, try case-insensitive name search
-    if not project:
-        project = await project_repository.get_by_name_case_insensitive(data.identifier)
-        if project:
-            resolution_method = "name"  # pragma: no cover
-
-    if not project:
-        raise HTTPException(status_code=404, detail=f"Project not found: '{data.identifier}'")
-
-    return ProjectResolveResponse(
-        external_id=project.external_id,
-        project_id=project.id,
-        name=project.name,
-        permalink=generate_permalink(project.name),
-        path=normalize_project_path(project.path),
-        is_active=project.is_active if hasattr(project, "is_active") else True,
-        is_default=project.is_default or False,
-        resolution_method=resolution_method,
-    )
-
-
-@router.get("/{project_id}", response_model=ProjectItem)
-async def get_project_by_id(
-    project_repository: ProjectRepositoryDep,
-    project_id: str = Path(..., description="Project external ID (UUID)"),
-) -> ProjectItem:
-    """Get project by its external ID (UUID).
-
-    This is the primary project retrieval method in v2, using stable UUID
-    identifiers that won't change with project renames.
-
-    Args:
-        project_id: External ID (UUID string)
-
-    Returns:
-        Project information including external_id
-
-    Raises:
-        HTTPException: 404 if project not found
-
-    Example:
-        GET /v2/projects/550e8400-e29b-41d4-a716-446655440000
-    """
-    logger.info(f"API v2 request: get_project_by_id for project_id={project_id}")
-
-    project = await project_repository.get_by_external_id(project_id)
-    if not project:
-        raise HTTPException(
-            status_code=404, detail=f"Project with external_id '{project_id}' not found"
-        )
-
-    return ProjectItem(
-        id=project.id,
-        external_id=project.external_id,
-        name=project.name,
-        path=normalize_project_path(project.path),
-        is_default=project.is_default or False,
-    )
-
-
-@router.get("/{project_id}/info", response_model=ProjectInfoResponse)
-async def get_project_info_by_id(
-    project_service: ProjectServiceDep,
-    project_repository: ProjectRepositoryDep,
-    project_id: str = Path(..., description="Project external ID (UUID)"),
-) -> ProjectInfoResponse:
-    """Get detailed project information by external ID."""
-    logger.info(f"API v2 request: get_project_info_by_id for project_id={project_id}")
-    project = await project_repository.get_by_external_id(project_id)
-    if not project:
-        raise HTTPException(
-            status_code=404, detail=f"Project with external_id '{project_id}' not found"
-        )
-    return await project_service.get_project_info(project.name)
-
-
-@router.patch("/{project_id}", response_model=ProjectStatusResponse)
-async def update_project_by_id(
-    project_service: ProjectServiceDep,
-    project_repository: ProjectRepositoryDep,
-    project_id: str = Path(..., description="Project external ID (UUID)"),
-    path: Optional[str] = Body(None, description="New absolute path for the project"),
-    is_active: Optional[bool] = Body(None, description="Status of the project (active/inactive)"),
-) -> ProjectStatusResponse:
-    """Update a project's information by external ID.
-
-    Args:
-        project_id: External ID (UUID string)
-        path: Optional new absolute path for the project
-        is_active: Optional status update for the project
-
-    Returns:
-        Response confirming the project was updated
-
-    Raises:
-        HTTPException: 400 if validation fails, 404 if project not found
-
-    Example:
-        PATCH /v2/projects/550e8400-e29b-41d4-a716-446655440000
-        {"path": "/new/path"}
-    """
-    logger.info(f"API v2 request: update_project_by_id for project_id={project_id}")
-
-    try:
-        # Validate that path is absolute if provided
-        if path and not os.path.isabs(path):
-            raise HTTPException(status_code=400, detail="Path must be absolute")
-
-        # Get original project info for the response
-        old_project = await project_repository.get_by_external_id(project_id)
-        if not old_project:
-            raise HTTPException(
-                status_code=404, detail=f"Project with external_id '{project_id}' not found"
-            )
-
-        old_project_info = ProjectItem(
-            id=old_project.id,
-            external_id=old_project.external_id,
-            name=old_project.name,
-            path=old_project.path,
-            is_default=old_project.is_default or False,
-        )
-
-        # Update using project name (service layer still uses names internally)
-        if path:
-            await project_service.move_project(old_project.name, path)
-        elif is_active is not None:
-            await project_service.update_project(old_project.name, is_active=is_active)
-
-        # Get updated project info (use the same external_id)
-        updated_project = await project_repository.get_by_external_id(project_id)
-        if not updated_project:  # pragma: no cover
-            raise HTTPException(
-                status_code=404,
-                detail=f"Project with external_id '{project_id}' not found after update",
-            )
-
-        return ProjectStatusResponse(
-            message=f"Project '{updated_project.name}' updated successfully",
-            status="success",
-            default=old_project.is_default or False,
-            old_project=old_project_info,
-            new_project=ProjectItem(
-                id=updated_project.id,
-                external_id=updated_project.external_id,
-                name=updated_project.name,
-                path=updated_project.path,
-                is_default=updated_project.is_default or False,
-            ),
-        )
-    except ValueError as e:  # pragma: no cover
-        raise HTTPException(status_code=400, detail=str(e))  # pragma: no cover
-
-
-@router.delete("/{project_id}", response_model=ProjectStatusResponse)
-async def delete_project_by_id(
-    project_service: ProjectServiceDep,
-    project_repository: ProjectRepositoryDep,
-    project_id: str = Path(..., description="Project external ID (UUID)"),
-    delete_notes: bool = Query(
-        False, description="If True, delete project directory from filesystem"
-    ),
-) -> ProjectStatusResponse:
-    """Delete a project by external ID.
-
-    Args:
-        project_id: External ID (UUID string)
-        delete_notes: If True, delete the project directory from the filesystem
-
-    Returns:
-        Response confirming the project was deleted
-
-    Raises:
-        HTTPException: 400 if trying to delete default project, 404 if not found
-
-    Example:
-        DELETE /v2/projects/550e8400-e29b-41d4-a716-446655440000?delete_notes=false
-    """
-    logger.info(
-        f"API v2 request: delete_project_by_id for project_id={project_id}, delete_notes={delete_notes}"
-    )
-
-    try:
-        old_project = await project_repository.get_by_external_id(project_id)
-        if not old_project:
-            raise HTTPException(
-                status_code=404, detail=f"Project with external_id '{project_id}' not found"
-            )
-
-        # Check if trying to delete the default project
-        # Use is_default from database, not ConfigManager (which doesn't work in cloud mode)
-        if old_project.is_default:
-            available_projects = await project_service.list_projects()
-            other_projects = [p.name for p in available_projects if p.external_id != project_id]
-            detail = f"Cannot delete default project '{old_project.name}'. "
-            if other_projects:
-                detail += (  # pragma: no cover
-                    f"Set another project as default first. Available: {', '.join(other_projects)}"
-                )
-            else:
-                detail += "This is the only project in your configuration."  # pragma: no cover
-            raise HTTPException(status_code=400, detail=detail)
-
-        # Delete using project name (service layer still uses names internally)
-        await project_service.remove_project(old_project.name, delete_notes=delete_notes)
-
-        return ProjectStatusResponse(
-            message=f"Project '{old_project.name}' removed successfully",
-            status="success",
-            default=False,
-            old_project=ProjectItem(
-                id=old_project.id,
-                external_id=old_project.external_id,
-                name=old_project.name,
-                path=old_project.path,
-                is_default=old_project.is_default or False,
-            ),
-            new_project=None,
-        )
-    except ValueError as e:  # pragma: no cover
-        raise HTTPException(status_code=400, detail=str(e))  # pragma: no cover
-
-
-@router.put("/{project_id}/default", response_model=ProjectStatusResponse)
-async def set_default_project_by_id(
-    project_service: ProjectServiceDep,
-    project_repository: ProjectRepositoryDep,
-    project_id: str = Path(..., description="Project external ID (UUID)"),
-) -> ProjectStatusResponse:
-    """Set a project as the default project by external ID.
-
-    Args:
-        project_id: External ID (UUID string) to set as default
-
-    Returns:
-        Response confirming the project was set as default
-
-    Raises:
-        HTTPException: 404 if project not found
-
-    Example:
-        PUT /v2/projects/550e8400-e29b-41d4-a716-446655440000/default
-    """
-    logger.info(f"API v2 request: set_default_project_by_id for project_id={project_id}")
-
-    try:
-        # Get the old default project from database
-        default_project = await project_repository.get_default_project()
-        if not default_project:
-            raise HTTPException(  # pragma: no cover
-                status_code=404, detail="No default project is currently set"
-            )
-
-        # Get the new default project by external_id
-        new_default_project = await project_repository.get_by_external_id(project_id)
-        if not new_default_project:
-            raise HTTPException(
-                status_code=404, detail=f"Project with external_id '{project_id}' not found"
-            )
-
-        # Set as default using project name (service layer still uses names internally)
-        await project_service.set_default_project(new_default_project.name)
-
-        return ProjectStatusResponse(
-            message=f"Project '{new_default_project.name}' set as default successfully",
-            status="success",
-            default=True,
-            old_project=ProjectItem(
-                id=default_project.id,
-                external_id=default_project.external_id,
-                name=default_project.name,
-                path=default_project.path,
-                is_default=False,
-            ),
-            new_project=ProjectItem(
-                id=new_default_project.id,
-                external_id=new_default_project.external_id,
-                name=new_default_project.name,
-                path=new_default_project.path,
-                is_default=True,
-            ),
-        )
-    except ValueError as e:  # pragma: no cover
-        raise HTTPException(status_code=400, detail=str(e))  # pragma: no cover

src/basic_memory/api/v2/routers/resource_router.py

@@ -1,289 +0,0 @@
-"""V2 Resource Router - ID-based resource content operations.
-
-This router uses entity external_ids (UUIDs) for all operations, with file paths
-in request bodies when needed. This is consistent with v2's external_id-first design.
-
-Key differences from v1:
-- Uses UUID external_ids in URL paths instead of integer IDs or file paths
-- File paths are in request bodies for create/update operations
-- More RESTful: POST for create, PUT for update, GET for read
-"""
-
-import uuid
-from pathlib import Path as PathLib
-
-from fastapi import APIRouter, HTTPException, Response, Path
-from loguru import logger
-
-from basic_memory.deps import (
-    ProjectConfigV2ExternalDep,
-    FileServiceV2ExternalDep,
-    EntityRepositoryV2ExternalDep,
-    SearchServiceV2ExternalDep,
-)
-from basic_memory.models.knowledge import Entity as EntityModel
-from basic_memory.schemas.v2.resource import (
-    CreateResourceRequest,
-    UpdateResourceRequest,
-    ResourceResponse,
-)
-from basic_memory.utils import validate_project_path
-
-router = APIRouter(prefix="/resource", tags=["resources-v2"])
-
-
-@router.get("/{entity_id}")
-async def get_resource_content(
-    config: ProjectConfigV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    file_service: FileServiceV2ExternalDep,
-    project_id: str = Path(..., description="Project external UUID"),
-    entity_id: str = Path(..., description="Entity external UUID"),
-) -> Response:
-    """Get raw resource content by entity external_id.
-
-    Args:
-        project_id: Project external UUID from URL path
-        entity_id: Entity external UUID
-        config: Project configuration
-        entity_repository: Entity repository for fetching entity data
-        file_service: File service for reading file content
-
-    Returns:
-        Response with entity content
-
-    Raises:
-        HTTPException: 404 if entity or file not found
-    """
-    logger.debug(f"V2 Getting content for project {project_id}, entity_id: {entity_id}")
-
-    # Get entity by external_id
-    entity = await entity_repository.get_by_external_id(entity_id)
-    if not entity:
-        raise HTTPException(status_code=404, detail=f"Entity {entity_id} not found")
-
-    # Validate entity file path to prevent path traversal
-    project_path = PathLib(config.home)
-    if not validate_project_path(entity.file_path, project_path):
-        logger.error(  # pragma: no cover
-            f"Invalid file path in entity {entity.id}: {entity.file_path}"
-        )
-        raise HTTPException(  # pragma: no cover
-            status_code=500,
-            detail="Entity contains invalid file path",
-        )
-
-    # Check file exists via file_service (for cloud compatibility)
-    if not await file_service.exists(entity.file_path):
-        raise HTTPException(  # pragma: no cover
-            status_code=404,
-            detail=f"File not found: {entity.file_path}",
-        )
-
-    # Read content via file_service as bytes (works with both local and S3)
-    content = await file_service.read_file_bytes(entity.file_path)
-    content_type = file_service.content_type(entity.file_path)
-
-    return Response(content=content, media_type=content_type)
-
-
-@router.post("", response_model=ResourceResponse)
-async def create_resource(
-    data: CreateResourceRequest,
-    config: ProjectConfigV2ExternalDep,
-    file_service: FileServiceV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    search_service: SearchServiceV2ExternalDep,
-    project_id: str = Path(..., description="Project external UUID"),
-) -> ResourceResponse:
-    """Create a new resource file.
-
-    Args:
-        project_id: Project external UUID from URL path
-        data: Create resource request with file_path and content
-        config: Project configuration
-        file_service: File service for writing files
-        entity_repository: Entity repository for creating entities
-        search_service: Search service for indexing
-
-    Returns:
-        ResourceResponse with file information including entity_id and external_id
-
-    Raises:
-        HTTPException: 400 for invalid file paths, 409 if file already exists
-    """
-    try:
-        # Validate path to prevent path traversal attacks
-        project_path = PathLib(config.home)
-        if not validate_project_path(data.file_path, project_path):
-            logger.warning(
-                f"Invalid file path attempted: {data.file_path} in project {config.name}"
-            )
-            raise HTTPException(
-                status_code=400,
-                detail=f"Invalid file path: {data.file_path}. "
-                "Path must be relative and stay within project boundaries.",
-            )
-
-        # Check if entity already exists
-        existing_entity = await entity_repository.get_by_file_path(data.file_path)
-        if existing_entity:
-            raise HTTPException(
-                status_code=409,
-                detail=f"Resource already exists at {data.file_path} with entity_id {existing_entity.external_id}. "
-                f"Use PUT /resource/{existing_entity.external_id} to update it.",
-            )
-
-        # Cloud compatibility: avoid assuming a local filesystem path.
-        # Delegate directory creation + writes to FileService (local or S3).
-        await file_service.ensure_directory(PathLib(data.file_path).parent)
-        checksum = await file_service.write_file(data.file_path, data.content)
-
-        # Get file info
-        file_metadata = await file_service.get_file_metadata(data.file_path)
-
-        # Determine file details
-        file_name = PathLib(data.file_path).name
-        content_type = file_service.content_type(data.file_path)
-        note_type = "canvas" if data.file_path.endswith(".canvas") else "file"
-
-        # Create a new entity model
-        # Explicitly set external_id to ensure NOT NULL constraint is satisfied (fixes #512)
-        entity = EntityModel(
-            external_id=str(uuid.uuid4()),
-            title=file_name,
-            note_type=note_type,
-            content_type=content_type,
-            file_path=data.file_path,
-            checksum=checksum,
-            created_at=file_metadata.created_at,
-            updated_at=file_metadata.modified_at,
-        )
-        entity = await entity_repository.add(entity)
-
-        # Index the file for search
-        await search_service.index_entity(entity)  # pyright: ignore
-
-        # Return success response
-        return ResourceResponse(
-            entity_id=entity.id,
-            external_id=entity.external_id,
-            file_path=data.file_path,
-            checksum=checksum,
-            size=file_metadata.size,
-            created_at=file_metadata.created_at.timestamp(),
-            modified_at=file_metadata.modified_at.timestamp(),
-        )
-    except HTTPException:
-        # Re-raise HTTP exceptions without wrapping
-        raise
-    except Exception as e:  # pragma: no cover
-        logger.error(f"Error creating resource {data.file_path}: {e}")
-        raise HTTPException(status_code=500, detail=f"Failed to create resource: {str(e)}")
-
-
-@router.put("/{entity_id}", response_model=ResourceResponse)
-async def update_resource(
-    data: UpdateResourceRequest,
-    config: ProjectConfigV2ExternalDep,
-    file_service: FileServiceV2ExternalDep,
-    entity_repository: EntityRepositoryV2ExternalDep,
-    search_service: SearchServiceV2ExternalDep,
-    project_id: str = Path(..., description="Project external UUID"),
-    entity_id: str = Path(..., description="Entity external UUID"),
-) -> ResourceResponse:
-    """Update an existing resource by entity external_id.
-
-    Can update content and optionally move the file to a new path.
-
-    Args:
-        project_id: Project external UUID from URL path
-        entity_id: Entity external UUID of the resource to update
-        data: Update resource request with content and optional new file_path
-        config: Project configuration
-        file_service: File service for writing files
-        entity_repository: Entity repository for updating entities
-        search_service: Search service for indexing
-
-    Returns:
-        ResourceResponse with updated file information
-
-    Raises:
-        HTTPException: 404 if entity not found, 400 for invalid paths
-    """
-    try:
-        # Get existing entity by external_id
-        entity = await entity_repository.get_by_external_id(entity_id)
-        if not entity:
-            raise HTTPException(status_code=404, detail=f"Entity {entity_id} not found")
-
-        # Determine target file path
-        target_file_path = data.file_path if data.file_path else entity.file_path
-
-        # Validate path to prevent path traversal attacks
-        project_path = PathLib(config.home)
-        if not validate_project_path(target_file_path, project_path):
-            logger.warning(
-                f"Invalid file path attempted: {target_file_path} in project {config.name}"
-            )
-            raise HTTPException(
-                status_code=400,
-                detail=f"Invalid file path: {target_file_path}. "
-                "Path must be relative and stay within project boundaries.",
-            )
-
-        # If moving file, handle the move
-        if data.file_path and data.file_path != entity.file_path:
-            # Ensure new parent directory exists (no-op for S3)
-            await file_service.ensure_directory(PathLib(target_file_path).parent)
-
-            # If old file exists, remove it via file_service (for cloud compatibility)
-            if await file_service.exists(entity.file_path):
-                await file_service.delete_file(entity.file_path)
-        else:
-            # Ensure directory exists for in-place update
-            await file_service.ensure_directory(PathLib(target_file_path).parent)
-
-        # Write content to target file
-        checksum = await file_service.write_file(target_file_path, data.content)
-
-        # Get file info
-        file_metadata = await file_service.get_file_metadata(target_file_path)
-
-        # Determine file details
-        file_name = PathLib(target_file_path).name
-        content_type = file_service.content_type(target_file_path)
-        note_type = "canvas" if target_file_path.endswith(".canvas") else "file"
-
-        # Update entity using internal ID
-        updated_entity = await entity_repository.update(
-            entity.id,
-            {
-                "title": file_name,
-                "note_type": note_type,
-                "content_type": content_type,
-                "file_path": target_file_path,
-                "checksum": checksum,
-                "updated_at": file_metadata.modified_at,
-            },
-        )
-
-        # Index the updated file for search
-        await search_service.index_entity(updated_entity)  # pyright: ignore
-
-        # Return success response
-        return ResourceResponse(
-            entity_id=entity.id,
-            external_id=entity.external_id,
-            file_path=target_file_path,
-            checksum=checksum,
-            size=file_metadata.size,
-            created_at=file_metadata.created_at.timestamp(),
-            modified_at=file_metadata.modified_at.timestamp(),
-        )
-    except HTTPException:
-        # Re-raise HTTP exceptions without wrapping
-        raise
-    except Exception as e:  # pragma: no cover
-        logger.error(f"Error updating resource {entity_id}: {e}")
-        raise HTTPException(status_code=500, detail=f"Failed to update resource: {str(e)}")

src/basic_memory/api/v2/routers/schema_router.py

@@ -1,412 +0,0 @@
-"""V2 router for schema operations.
-
-Provides endpoints for schema validation, inference, and drift detection.
-The schema system validates notes against Picoschema definitions without
-introducing any new data model -- it works entirely with existing
-observations and relations.
-
-Flow: Entity loaded with eager observations/relations -> convert to tuples -> core functions.
-"""
-
-from pathlib import Path as FilePath
-
-import frontmatter
-from fastapi import APIRouter, Path, Query
-from loguru import logger
-
-from basic_memory.deps import (
-    EntityRepositoryV2ExternalDep,
-    FileServiceV2ExternalDep,
-    LinkResolverV2ExternalDep,
-)
-from basic_memory.models.knowledge import Entity
-from basic_memory.schemas.schema import (
-    ValidationReport,
-    InferenceReport,
-    DriftReport,
-    NoteValidationResponse,
-    FieldResultResponse,
-    FieldFrequencyResponse,
-    DriftFieldResponse,
-)
-from basic_memory.schema.resolver import resolve_schema
-from basic_memory.schema.validator import validate_note
-from basic_memory.schema.inference import infer_schema, NoteData, ObservationData, RelationData
-from basic_memory.schema.diff import diff_schema
-from basic_memory.utils import generate_permalink
-
-# Note: No prefix here -- it's added during registration as /v2/{project_id}/schema
-router = APIRouter(tags=["schema"])
-
-
-# --- ORM to core data conversion ---
-
-
-def _entity_observations(entity: Entity) -> list[ObservationData]:
-    """Extract ObservationData from an entity's observations."""
-    return [ObservationData(obs.category, obs.content) for obs in entity.observations]
-
-
-def _entity_relations(entity: Entity) -> list[RelationData]:
-    """Extract RelationData from an entity's outgoing relations.
-
-    Carries the target entity's type on each relation so the inference engine
-    can suggest correct types (e.g. works_at -> Organization, not the source type).
-    """
-    return [
-        RelationData(
-            relation_type=rel.relation_type,
-            target_name=rel.to_name,
-            target_note_type=rel.to_entity.note_type if rel.to_entity else None,
-        )
-        for rel in entity.outgoing_relations
-    ]
-
-
-def _entity_to_note_data(entity: Entity) -> NoteData:
-    """Convert an ORM Entity to a NoteData for inference/diff analysis."""
-    return NoteData(
-        identifier=entity.permalink or entity.file_path,
-        observations=_entity_observations(entity),
-        relations=_entity_relations(entity),
-    )
-
-
-def _entity_frontmatter(entity: Entity) -> dict:
-    """Build a frontmatter dict from an entity's database metadata.
-
-    Used for the notes being validated — their type and schema ref are
-    unlikely to change between syncs.
-    """
-    fm = dict(entity.entity_metadata) if entity.entity_metadata else {}
-    if entity.note_type:
-        fm.setdefault("type", entity.note_type)
-    return fm
-
-
-async def _schema_frontmatter_from_file(
-    file_service: FileServiceV2ExternalDep,
-    entity: Entity,
-) -> dict:
-    """Read a schema entity's frontmatter directly from its file.
-
-    Schema definitions (field declarations, validation mode) are the source
-    of truth for validation. Reading from the file ensures schema-validate
-    always uses the latest settings, even when the file watcher hasn't
-    synced changes to entity_metadata in the database.
-    """
-    try:
-        content = await file_service.read_file_content(entity.file_path)
-        post = frontmatter.loads(content)
-        metadata = dict(post.metadata)
-
-        # Trigger: file is mid-edit and missing required schema fields
-        # Why: parse_schema_note() raises ValueError for missing entity/schema,
-        #   which would turn validation into a 500 response
-        # Outcome: fall back to last-known-good database metadata
-        if not metadata.get("entity") or not isinstance(metadata.get("schema"), dict):
-            logger.warning(
-                "Schema file has incomplete frontmatter, falling back to database metadata",
-                file_path=entity.file_path,
-            )
-            return _entity_frontmatter(entity)
-
-        return metadata
-    except Exception:
-        # Trigger: file is missing, unreadable, or has malformed frontmatter
-        # Why: fall back to database metadata rather than failing validation entirely
-        # Outcome: behaves like before this change — uses potentially stale data
-        logger.warning(
-            "Failed to read schema file, falling back to database metadata",
-            file_path=entity.file_path,
-        )
-        return _entity_frontmatter(entity)
-
-
-# --- Validation ---
-
-
-@router.post("/schema/validate", response_model=ValidationReport)
-async def validate_schema(
-    entity_repository: EntityRepositoryV2ExternalDep,
-    file_service: FileServiceV2ExternalDep,
-    link_resolver: LinkResolverV2ExternalDep,
-    project_id: str = Path(..., description="Project external UUID"),
-    note_type: str | None = Query(None, description="Note type to validate"),
-    identifier: str | None = Query(None, description="Specific note identifier"),
-):
-    """Validate notes against their resolved schemas.
-
-    Validates a specific note (by identifier) or all notes of a given type.
-    Returns warnings/errors based on the schema's validation mode.
-
-    Schema definitions are read directly from their files to ensure the
-    latest settings (validation mode, field declarations) are always used,
-    even when file changes haven't been synced to the database yet.
-    """
-    results: list[NoteValidationResponse] = []
-
-    # --- Single note validation ---
-    if identifier:
-        # Resolve identifier flexibly (permalink, title, path, fuzzy)
-        # to match how read_note and other tools resolve identifiers
-        entity = await link_resolver.resolve_link(identifier)
-        if not entity:
-            return ValidationReport(note_type=note_type, total_notes=0, total_entities=0)
-
-        frontmatter = _entity_frontmatter(entity)
-        schema_ref = frontmatter.get("schema")
-
-        async def search_fn(query: str) -> list[dict]:
-            entities = await _find_schema_entities(
-                entity_repository,
-                query,
-                allow_reference_match=isinstance(schema_ref, str) and query == schema_ref,
-            )
-            return [await _schema_frontmatter_from_file(file_service, e) for e in entities]
-
-        schema_def = await resolve_schema(frontmatter, search_fn)
-        if schema_def:
-            result = validate_note(
-                entity.title or entity.permalink or identifier,
-                schema_def,
-                _entity_observations(entity),
-                _entity_relations(entity),
-                frontmatter=frontmatter,
-            )
-            results.append(_to_note_validation_response(result))
-
-        return ValidationReport(
-            note_type=note_type or entity.note_type,
-            total_notes=len(results),
-            total_entities=1,
-            valid_count=1 if (results and results[0].passed) else 0,
-            warning_count=sum(len(r.warnings) for r in results),
-            error_count=sum(len(r.errors) for r in results),
-            results=results,
-        )
-
-    # --- Batch validation by note type ---
-    entities = await _find_by_note_type(entity_repository, note_type) if note_type else []
-
-    for entity in entities:
-        frontmatter = _entity_frontmatter(entity)
-        schema_ref = frontmatter.get("schema")
-
-        async def search_fn(query: str) -> list[dict]:
-            entities = await _find_schema_entities(
-                entity_repository,
-                query,
-                allow_reference_match=isinstance(schema_ref, str) and query == schema_ref,
-            )
-            return [await _schema_frontmatter_from_file(file_service, e) for e in entities]
-
-        schema_def = await resolve_schema(frontmatter, search_fn)
-        if schema_def:
-            result = validate_note(
-                entity.title or entity.permalink or entity.file_path,
-                schema_def,
-                _entity_observations(entity),
-                _entity_relations(entity),
-                frontmatter=frontmatter,
-            )
-            results.append(_to_note_validation_response(result))
-
-    valid = sum(1 for r in results if r.passed)
-    return ValidationReport(
-        note_type=note_type,
-        total_notes=len(results),
-        total_entities=len(entities),
-        valid_count=valid,
-        warning_count=sum(len(r.warnings) for r in results),
-        error_count=sum(len(r.errors) for r in results),
-        results=results,
-    )
-
-
-# --- Inference ---
-
-
-@router.post("/schema/infer", response_model=InferenceReport)
-async def infer_schema_endpoint(
-    entity_repository: EntityRepositoryV2ExternalDep,
-    project_id: str = Path(..., description="Project external UUID"),
-    note_type: str = Query(..., description="Note type to analyze"),
-    threshold: float = Query(0.25, description="Minimum frequency for optional fields"),
-):
-    """Infer a schema from existing notes of a given type.
-
-    Examines observation categories and relation types across all notes
-    of the given type. Returns frequency analysis and suggested Picoschema.
-    """
-    entities = await _find_by_note_type(entity_repository, note_type)
-    notes_data = [_entity_to_note_data(entity) for entity in entities]
-
-    result = infer_schema(note_type, notes_data, optional_threshold=threshold)
-
-    return InferenceReport(
-        note_type=result.note_type,
-        notes_analyzed=result.notes_analyzed,
-        field_frequencies=[
-            FieldFrequencyResponse(
-                name=f.name,
-                source=f.source,
-                count=f.count,
-                total=f.total,
-                percentage=f.percentage,
-                sample_values=f.sample_values,
-                is_array=f.is_array,
-                target_type=f.target_type,
-            )
-            for f in result.field_frequencies
-        ],
-        suggested_schema=result.suggested_schema,
-        suggested_required=result.suggested_required,
-        suggested_optional=result.suggested_optional,
-        excluded=result.excluded,
-    )
-
-
-# --- Drift Detection ---
-
-
-@router.get("/schema/diff/{note_type}", response_model=DriftReport)
-async def diff_schema_endpoint(
-    entity_repository: EntityRepositoryV2ExternalDep,
-    file_service: FileServiceV2ExternalDep,
-    note_type: str = Path(..., description="Note type to check for drift"),
-    project_id: str = Path(..., description="Project external UUID"),
-):
-    """Show drift between a schema definition and actual note usage.
-
-    Compares the existing schema for an entity type against how notes
-    of that type are actually structured. Identifies new fields, dropped
-    fields, and cardinality changes.
-    """
-
-    async def search_fn(query: str) -> list[dict]:
-        entities = await _find_schema_entities(entity_repository, query)
-        return [await _schema_frontmatter_from_file(file_service, e) for e in entities]
-
-    # Resolve schema by note type
-    schema_frontmatter = {"type": note_type}
-    schema_def = await resolve_schema(schema_frontmatter, search_fn)
-
-    if not schema_def:
-        return DriftReport(note_type=note_type, schema_found=False)
-
-    # Collect all notes of this type
-    entities = await _find_by_note_type(entity_repository, note_type)
-    notes_data = [_entity_to_note_data(entity) for entity in entities]
-
-    result = diff_schema(schema_def, notes_data)
-
-    return DriftReport(
-        note_type=note_type,
-        new_fields=[
-            DriftFieldResponse(
-                name=f.name,
-                source=f.source,
-                count=f.count,
-                total=f.total,
-                percentage=f.percentage,
-            )
-            for f in result.new_fields
-        ],
-        dropped_fields=[
-            DriftFieldResponse(
-                name=f.name,
-                source=f.source,
-                count=f.count,
-                total=f.total,
-                percentage=f.percentage,
-            )
-            for f in result.dropped_fields
-        ],
-        cardinality_changes=result.cardinality_changes,
-    )
-
-
-# --- Helpers ---
-
-
-async def _find_by_note_type(
-    entity_repository: EntityRepositoryV2ExternalDep,
-    note_type: str,
-) -> list[Entity]:
-    """Find all entities of a given type using the repository's select pattern."""
-    query = entity_repository.select().where(Entity.note_type == note_type)
-    result = await entity_repository.execute_query(query)
-    return list(result.scalars().all())
-
-
-async def _find_schema_entities(
-    entity_repository: EntityRepositoryV2ExternalDep,
-    target_note_type: str,
-    *,
-    allow_reference_match: bool = False,
-) -> list[Entity]:
-    """Find schema entities for resolver lookups.
-
-    Resolution strategy:
-    1) Always try exact entity_metadata['entity'] match (for implicit type lookup
-       and explicit references that use entity names)
-    2) Only when allow_reference_match=True and no entity match was found, try
-       exact reference matching by title/permalink (explicit schema references)
-    """
-    query = entity_repository.select().where(Entity.note_type == "schema")
-    result = await entity_repository.execute_query(query)
-    entities = list(result.scalars().all())
-
-    normalized_target = generate_permalink(target_note_type)
-
-    entity_matches = [
-        e
-        for e in entities
-        if e.entity_metadata
-        and isinstance(e.entity_metadata.get("entity"), str)
-        and generate_permalink(e.entity_metadata["entity"]) == normalized_target
-    ]
-    if entity_matches:
-        return entity_matches
-
-    if not allow_reference_match:
-        return []
-
-    reference_matches: list[Entity] = []
-    for entity in entities:
-        candidate_refs: list[str] = []
-        if entity.title:
-            candidate_refs.append(entity.title)
-        if entity.permalink:
-            candidate_refs.append(entity.permalink)
-            candidate_refs.append(FilePath(entity.permalink).name)
-
-        if any(generate_permalink(ref) == normalized_target for ref in candidate_refs):
-            reference_matches.append(entity)
-
-    return reference_matches
-
-
-def _to_note_validation_response(result) -> NoteValidationResponse:
-    """Convert a core ValidationResult to a Pydantic response model."""
-    return NoteValidationResponse(
-        note_identifier=result.note_identifier,
-        schema_entity=result.schema_entity,
-        passed=result.passed,
-        field_results=[
-            FieldResultResponse(
-                field_name=fr.field.name,
-                field_type=fr.field.type,
-                required=fr.field.required,
-                status=fr.status,
-                values=fr.values,
-                message=fr.message,
-            )
-            for fr in result.field_results
-        ],
-        unmatched_observations=result.unmatched_observations,
-        unmatched_relations=result.unmatched_relations,
-        warnings=result.warnings,
-        errors=result.errors,
-    )

src/basic_memory/api/v2/routers/search_router.py

@@ -1,94 +0,0 @@
-"""V2 router for search operations.
-
-This router uses external_id UUIDs for stable, API-friendly routing.
-V1 uses string-based project names which are less efficient and less stable.
-"""
-
-from fastapi import APIRouter, HTTPException, Path
-
-from basic_memory.api.v2.utils import to_search_results
-from basic_memory.repository.semantic_errors import (
-    SemanticDependenciesMissingError,
-    SemanticSearchDisabledError,
-)
-from basic_memory.schemas.search import SearchQuery, SearchResponse
-from basic_memory.deps import (
-    SearchServiceV2ExternalDep,
-    EntityServiceV2ExternalDep,
-    TaskSchedulerDep,
-    ProjectExternalIdPathDep,
-)
-
-# Note: No prefix here - it's added during registration as /v2/{project_id}/search
-router = APIRouter(tags=["search"])
-
-
-@router.post("/search/", response_model=SearchResponse)
-async def search(
-    query: SearchQuery,
-    search_service: SearchServiceV2ExternalDep,
-    entity_service: EntityServiceV2ExternalDep,
-    project_id: str = Path(..., description="Project external UUID"),
-    page: int = 1,
-    page_size: int = 10,
-):
-    """Search across all knowledge and documents in a project.
-
-    V2 uses external_id UUIDs for stable API references.
-
-    Args:
-        project_id: Project external UUID from URL path
-        query: Search query parameters (text, filters, etc.)
-        search_service: Search service scoped to project
-        entity_service: Entity service scoped to project
-        page: Page number for pagination
-        page_size: Number of results per page
-
-    Returns:
-        SearchResponse with paginated search results
-    """
-    offset = (page - 1) * page_size
-    # Fetch one extra item to detect whether more pages exist (N+1 trick)
-    fetch_limit = page_size + 1
-    try:
-        results = await search_service.search(query, limit=fetch_limit, offset=offset)
-    except SemanticSearchDisabledError as exc:
-        raise HTTPException(status_code=400, detail=str(exc)) from exc
-    except SemanticDependenciesMissingError as exc:
-        raise HTTPException(status_code=400, detail=str(exc)) from exc
-    except ValueError as exc:
-        raise HTTPException(status_code=400, detail=str(exc)) from exc
-
-    has_more = len(results) > page_size
-    if has_more:
-        results = results[:page_size]
-
-    search_results = await to_search_results(entity_service, results)
-    return SearchResponse(
-        results=search_results,
-        current_page=page,
-        page_size=page_size,
-        has_more=has_more,
-    )
-
-
-@router.post("/search/reindex")
-async def reindex(
-    task_scheduler: TaskSchedulerDep,
-    project_id: ProjectExternalIdPathDep,
-):
-    """Recreate and populate the search index for a project.
-
-    This is a background operation that rebuilds the search index
-    from scratch. Useful after bulk updates or if the index becomes
-    corrupted.
-
-    Args:
-        project_id: Project external UUID from URL path
-        task_scheduler: Task scheduler for background work
-
-    Returns:
-        Status message indicating reindex has been initiated
-    """
-    task_scheduler.schedule("reindex_project", project_id=project_id)
-    return {"status": "ok", "message": "Reindex initiated"}

src/basic_memory/cli/analytics.py

@@ -1,114 +0,0 @@
-"""Lightweight CLI analytics via Umami event collector.
-
-Sends anonymous, non-blocking usage events to help understand how the
-CLI-to-cloud conversion funnel performs. No PII, no fingerprinting,
-no cookies. Respects the same opt-out mechanisms as promo messaging.
-
-Events are fire-and-forget — analytics never blocks or breaks the CLI.
-
-Defaults point to the Basic Memory Umami Cloud instance. Override via:
-    BASIC_MEMORY_UMAMI_HOST     — Custom Umami instance URL
-    BASIC_MEMORY_UMAMI_SITE_ID  — Custom Website ID
-Opt out entirely with BASIC_MEMORY_NO_PROMOS=1.
-"""
-
-import json
-import os
-import threading
-import urllib.request
-from typing import Optional
-
-import basic_memory
-
-
-# ---------------------------------------------------------------------------
-# Configuration — defaults baked in, overridable via environment
-# ---------------------------------------------------------------------------
-
-_DEFAULT_UMAMI_HOST = "https://api-gateway.umami.dev"
-_DEFAULT_UMAMI_SITE_ID = "f6479898-ebaf-4e60-bce2-6dc60a3f6c5c"
-
-
-def _umami_host() -> Optional[str]:
-    return os.getenv("BASIC_MEMORY_UMAMI_HOST", "").strip() or _DEFAULT_UMAMI_HOST
-
-
-def _umami_site_id() -> Optional[str]:
-    return os.getenv("BASIC_MEMORY_UMAMI_SITE_ID", "").strip() or _DEFAULT_UMAMI_SITE_ID
-
-
-def _analytics_disabled() -> bool:
-    """True when analytics should not fire."""
-    value = os.getenv("BASIC_MEMORY_NO_PROMOS", "").strip().lower()
-    return value in {"1", "true", "yes"}
-
-
-def _is_configured() -> bool:
-    """True when both host and site ID are available."""
-    return _umami_host() is not None and _umami_site_id() is not None
-
-
-# ---------------------------------------------------------------------------
-# Public API
-# ---------------------------------------------------------------------------
-
-# Well-known event names for the promo/cloud funnel
-EVENT_PROMO_SHOWN = "cli-promo-shown"
-EVENT_PROMO_OPTED_OUT = "cli-promo-opted-out"
-EVENT_CLOUD_LOGIN_STARTED = "cli-cloud-login-started"
-EVENT_CLOUD_LOGIN_SUCCESS = "cli-cloud-login-success"
-EVENT_CLOUD_LOGIN_SUB_REQUIRED = "cli-cloud-login-sub-required"
-
-
-def track(event_name: str, data: Optional[dict] = None) -> None:
-    """Send an analytics event to Umami. Non-blocking, silent on failure.
-
-    Parameters
-    ----------
-    event_name:
-        Short kebab-case name (e.g. "cli-promo-shown").
-    data:
-        Optional dict of event properties (all values should be strings/numbers).
-    """
-    if _analytics_disabled() or not _is_configured():
-        return
-
-    host = _umami_host()
-    site_id = _umami_site_id()
-
-    # Umami v2 /api/send requires "type" at top level alongside "payload"
-    payload = {
-        "type": "event",
-        "payload": {
-            "hostname": "cli.basicmemory.com",
-            "language": "en",
-            "url": f"/cli/{event_name}",
-            "website": site_id,
-            "name": event_name,
-            "data": {
-                "version": basic_memory.__version__,
-                **(data or {}),
-            },
-        },
-    }
-
-    def _send():
-        try:
-            req = urllib.request.Request(
-                f"{host}/api/send",
-                data=json.dumps(payload).encode("utf-8"),
-                headers={
-                    "Content-Type": "application/json",
-                    # Umami's bot detection rejects non-browser User-Agents
-                    "User-Agent": "Mozilla/5.0 (compatible; BasicMemoryCLI/"
-                    f"{basic_memory.__version__})",
-                },
-            )
-            urllib.request.urlopen(req, timeout=3)
-        except Exception:
-            pass  # Never break the CLI for analytics
-
-    # Non-daemon so the process waits for the request to complete.
-    # The 3s urllib timeout caps the worst-case exit delay.
-    t = threading.Thread(target=_send)
-    t.start()

src/basic_memory/cli/app.py

@@ -1,25 +1,20 @@
-# This prevents DEBUG logs from appearing on stdout during module-level
-# initialization (e.g., template_loader.TemplateLoader() logs at DEBUG level).
-from loguru import logger
+from typing import Optional
 
-logger.remove()
+import typer
 
-from typing import Optional  # noqa: E402
-
-import typer  # noqa: E402
-
-from basic_memory.cli.auto_update import maybe_run_periodic_auto_update  # noqa: E402
-from basic_memory.cli.container import CliContainer, set_container  # noqa: E402
-from basic_memory.cli.promo import maybe_show_cloud_promo, maybe_show_init_line  # noqa: E402
-from basic_memory.config import init_cli_logging  # noqa: E402
+from basic_memory.config import get_project_config
+from basic_memory.mcp.project_session import session
 
 
 def version_callback(value: bool) -> None:
     """Show version and exit."""
     if value:  # pragma: no cover
         import basic_memory
+        from basic_memory.config import config
 
         typer.echo(f"Basic Memory version: {basic_memory.__version__}")
+        typer.echo(f"Current project: {config.project}")
+        typer.echo(f"Project path: {config.home}")
         raise typer.Exit()
 
 
@@ -29,6 +24,13 @@ app = typer.Typer(name="basic-memory")
 @app.callback()
 def app_callback(
     ctx: typer.Context,
+    project: Optional[str] = typer.Option(
+        None,
+        "--project",
+        "-p",
+        help="Specify which project to use 1",
+        envvar="BASIC_MEMORY_PROJECT",
+    ),
     version: Optional[bool] = typer.Option(
         None,
         "--version",
@@ -40,64 +42,32 @@ def app_callback(
 ) -> None:
     """Basic Memory - Local-first personal knowledge management."""
 
-    # Initialize logging for CLI (file only, no stdout)
-    init_cli_logging()
-
-    # --- Composition Root ---
-    # Create container and read config (single point of config access)
-    container = CliContainer.create()
-    set_container(container)
-
-    # Trigger: first-run init confirmation before command output.
-    # Why: informational "initialized" message belongs above command results, not in the upsell panel.
-    # Outcome: one-time plain line printed before the subcommand runs.
-    maybe_show_init_line(ctx.invoked_subcommand)
-
-    # Trigger: register post-command messaging callbacks.
-    # Why: informational/promo/update output belongs below command results.
-    # Outcome: command output remains primary, with optional follow-up notices afterwards.
-    def _post_command_messages() -> None:
-        maybe_show_cloud_promo(ctx.invoked_subcommand)
-        maybe_run_periodic_auto_update(ctx.invoked_subcommand)
-
-    ctx.call_on_close(_post_command_messages)
-
-    # Run initialization for commands that don't use the API
-    # Skip for 'mcp' command - it has its own lifespan that handles initialization
-    # Skip for API-using commands (status, sync, etc.) - they handle initialization via deps.py
-    # Skip for 'reset' command - it manages its own database lifecycle
-    skip_init_commands = {
-        "doctor",
-        "mcp",
-        "status",
-        "sync",
-        "project",
-        "tool",
-        "reset",
-        "reindex",
-        "update",
-        "watch",
-    }
-    if (
-        not version
-        and ctx.invoked_subcommand is not None
-        and ctx.invoked_subcommand not in skip_init_commands
-    ):
+    # Run initialization for every command unless --version was specified
+    if not version and ctx.invoked_subcommand is not None:
+        from basic_memory.config import app_config
         from basic_memory.services.initialization import ensure_initialization
 
-        ensure_initialization(container.config)
+        ensure_initialization(app_config)
+
+        # Initialize MCP session with the specified project or default
+        if project:  # pragma: no cover
+            # Use the project specified via --project flag
+            current_project_config = get_project_config(project)
+            session.set_current_project(current_project_config.name)
+
+            # Update the global config to use this project
+            from basic_memory.config import update_current_project
+
+            update_current_project(project)
+        else:
+            # Use the default project
+            current_project = app_config.default_project
+            session.set_current_project(current_project)
 
 
-## import
 # Register sub-command groups
 import_app = typer.Typer(help="Import data from various sources")
 app.add_typer(import_app, name="import")
 
-claude_app = typer.Typer(help="Import Conversations from Claude JSON export.")
+claude_app = typer.Typer()
 import_app.add_typer(claude_app, name="claude")
-
-
-## cloud
-
-cloud_app = typer.Typer(help="Access Basic Memory Cloud")
-app.add_typer(cloud_app, name="cloud")

src/basic_memory/cli/auth.py

@@ -1,300 +0,0 @@
-"""WorkOS OAuth Device Authorization for CLI."""
-
-import base64
-import hashlib
-import json
-import os
-import secrets
-import time
-import webbrowser
-from contextlib import asynccontextmanager
-from collections.abc import AsyncIterator, Callable
-from typing import AsyncContextManager
-
-import httpx
-from rich.console import Console
-
-from basic_memory.config import ConfigManager
-
-console = Console()
-
-
-class CLIAuth:
-    """Handles WorkOS OAuth Device Authorization for CLI tools."""
-
-    def __init__(
-        self,
-        client_id: str,
-        authkit_domain: str,
-        http_client_factory: Callable[[], AsyncContextManager[httpx.AsyncClient]] | None = None,
-    ):
-        self.client_id = client_id
-        self.authkit_domain = authkit_domain
-        app_config = ConfigManager().config
-        # Store tokens in data dir
-        self.token_file = app_config.data_dir_path / "basic-memory-cloud.json"
-        # PKCE parameters
-        self.code_verifier = None
-        self.code_challenge = None
-        self._http_client_factory = http_client_factory
-
-    @asynccontextmanager
-    async def _get_http_client(self) -> AsyncIterator[httpx.AsyncClient]:
-        """Create an AsyncClient, optionally via injected factory.
-
-        Why: enables reliable tests without monkeypatching httpx internals while
-        still using real httpx request/response objects.
-        """
-        if self._http_client_factory:
-            async with self._http_client_factory() as client:
-                yield client
-        else:
-            async with httpx.AsyncClient() as client:
-                yield client
-
-    def generate_pkce_pair(self) -> tuple[str, str]:
-        """Generate PKCE code verifier and challenge."""
-        # Generate code verifier (43-128 characters)
-        code_verifier = base64.urlsafe_b64encode(secrets.token_bytes(32)).decode("utf-8")
-        code_verifier = code_verifier.rstrip("=")
-
-        # Generate code challenge (SHA256 hash of verifier)
-        challenge_bytes = hashlib.sha256(code_verifier.encode("utf-8")).digest()
-        code_challenge = base64.urlsafe_b64encode(challenge_bytes).decode("utf-8")
-        code_challenge = code_challenge.rstrip("=")
-
-        return code_verifier, code_challenge
-
-    async def request_device_authorization(self) -> dict | None:
-        """Request device authorization from WorkOS with PKCE."""
-        device_auth_url = f"{self.authkit_domain}/oauth2/device_authorization"
-
-        # Generate PKCE pair
-        self.code_verifier, self.code_challenge = self.generate_pkce_pair()
-
-        data = {
-            "client_id": self.client_id,
-            "scope": "openid profile email offline_access",
-            "code_challenge": self.code_challenge,
-            "code_challenge_method": "S256",
-        }
-
-        try:
-            async with self._get_http_client() as client:
-                response = await client.post(device_auth_url, data=data)
-
-                if response.status_code == 200:
-                    return response.json()
-                else:
-                    console.print(
-                        f"[red]Device authorization failed: {response.status_code} - {response.text}[/red]"
-                    )
-                    return None
-        except Exception as e:
-            console.print(f"[red]Device authorization error: {e}[/red]")
-            return None
-
-    def display_user_instructions(self, device_response: dict) -> None:
-        """Display user instructions for device authorization."""
-        user_code = device_response["user_code"]
-        verification_uri = device_response["verification_uri"]
-        verification_uri_complete = device_response.get("verification_uri_complete")
-
-        console.print("\n[bold blue]Authentication Required[/bold blue]")
-        console.print("\nTo authenticate, please visit:")
-        console.print(f"[bold cyan]{verification_uri}[/bold cyan]")
-        console.print(f"\nAnd enter this code: [bold yellow]{user_code}[/bold yellow]")
-
-        if verification_uri_complete:
-            console.print("\nOr for one-click access, visit:")
-            console.print(f"[bold green]{verification_uri_complete}[/bold green]")
-
-            # Try to open browser automatically
-            try:
-                console.print("\n[dim]Opening browser automatically...[/dim]")
-                webbrowser.open(verification_uri_complete)
-            except Exception:
-                pass  # Silently fail if browser can't be opened
-
-        console.print("\n[dim]Waiting for you to complete authentication in your browser...[/dim]")
-
-    async def poll_for_token(self, device_code: str, interval: int = 5) -> dict | None:
-        """Poll the token endpoint until user completes authentication."""
-        token_url = f"{self.authkit_domain}/oauth2/token"
-
-        data = {
-            "client_id": self.client_id,
-            "device_code": device_code,
-            "grant_type": "urn:ietf:params:oauth:grant-type:device_code",
-            "code_verifier": self.code_verifier,
-        }
-
-        max_attempts = 60  # 5 minutes with 5-second intervals
-        current_interval = interval
-
-        for _attempt in range(max_attempts):
-            try:
-                async with self._get_http_client() as client:
-                    response = await client.post(token_url, data=data)
-
-                    if response.status_code == 200:
-                        return response.json()
-
-                    # Parse error response
-                    try:
-                        error_data = response.json()
-                        error = error_data.get("error")
-                    except Exception:
-                        error = "unknown_error"
-
-                    if error == "authorization_pending":
-                        # User hasn't completed auth yet, keep polling
-                        pass
-                    elif error == "slow_down":
-                        # Increase polling interval
-                        current_interval += 5
-                        console.print("[yellow]Slowing down polling rate...[/yellow]")
-                    elif error == "access_denied":
-                        console.print("[red]Authentication was denied by user[/red]")
-                        return None
-                    elif error == "expired_token":
-                        console.print("[red]Device code has expired. Please try again.[/red]")
-                        return None
-                    else:
-                        console.print(f"[red]Token polling error: {error}[/red]")
-                        return None
-
-            except Exception as e:
-                console.print(f"[red]Token polling request error: {e}[/red]")
-
-            # Wait before next poll
-            await self._async_sleep(current_interval)
-
-        console.print("[red]Authentication timeout. Please try again.[/red]")
-        return None
-
-    async def _async_sleep(self, seconds: int) -> None:
-        """Async sleep utility."""
-        import asyncio
-
-        await asyncio.sleep(seconds)
-
-    def save_tokens(self, tokens: dict) -> None:
-        """Save tokens to project root as .bm-auth.json."""
-        token_data = {
-            "access_token": tokens["access_token"],
-            "refresh_token": tokens.get("refresh_token"),
-            "expires_at": int(time.time()) + tokens.get("expires_in", 3600),
-            "token_type": tokens.get("token_type", "Bearer"),
-        }
-
-        with open(self.token_file, "w") as f:
-            json.dump(token_data, f, indent=2)
-
-        # Secure the token file
-        os.chmod(self.token_file, 0o600)
-
-        console.print(f"[green]Tokens saved to {self.token_file}[/green]")
-
-    def load_tokens(self) -> dict | None:
-        """Load tokens from .bm-auth.json file."""
-        if not self.token_file.exists():
-            return None
-
-        try:
-            with open(self.token_file) as f:
-                return json.load(f)
-        except (OSError, json.JSONDecodeError):
-            return None
-
-    def is_token_valid(self, tokens: dict) -> bool:
-        """Check if stored token is still valid."""
-        expires_at = tokens.get("expires_at", 0)
-        # Add 60 second buffer for clock skew
-        return time.time() < (expires_at - 60)
-
-    async def refresh_token(self, refresh_token: str) -> dict | None:
-        """Refresh access token using refresh token."""
-        token_url = f"{self.authkit_domain}/oauth2/token"
-
-        data = {
-            "client_id": self.client_id,
-            "grant_type": "refresh_token",
-            "refresh_token": refresh_token,
-        }
-
-        try:
-            async with self._get_http_client() as client:
-                response = await client.post(token_url, data=data)
-
-                if response.status_code == 200:
-                    return response.json()
-                else:
-                    console.print(
-                        f"[red]Token refresh failed: {response.status_code} - {response.text}[/red]"
-                    )
-                    return None
-        except Exception as e:
-            console.print(f"[red]Token refresh error: {e}[/red]")
-            return None
-
-    async def get_valid_token(self) -> str | None:
-        """Get valid access token, refresh if needed."""
-        tokens = self.load_tokens()
-        if not tokens:
-            return None
-
-        if self.is_token_valid(tokens):
-            return tokens["access_token"]
-
-        # Token expired - try to refresh if we have a refresh token
-        refresh_token = tokens.get("refresh_token")
-        if refresh_token:
-            console.print("[yellow]Access token expired, refreshing...[/yellow]")
-
-            new_tokens = await self.refresh_token(refresh_token)
-            if new_tokens:
-                # Save new tokens (may include rotated refresh token)
-                self.save_tokens(new_tokens)
-                console.print("[green]Token refreshed successfully[/green]")
-                return new_tokens["access_token"]
-            else:
-                console.print("[yellow]Token refresh failed. Please run 'login' again.[/yellow]")
-                return None
-        else:
-            console.print("[yellow]No refresh token available. Please run 'login' again.[/yellow]")
-            return None
-
-    async def login(self) -> bool:
-        """Perform OAuth Device Authorization login flow."""
-        console.print("[blue]Initiating authentication...[/blue]")
-
-        # Step 1: Request device authorization
-        device_response = await self.request_device_authorization()
-        if not device_response:
-            return False
-
-        # Step 2: Display user instructions
-        self.display_user_instructions(device_response)
-
-        # Step 3: Poll for token
-        device_code = device_response["device_code"]
-        interval = device_response.get("interval", 5)
-
-        tokens = await self.poll_for_token(device_code, interval)
-        if not tokens:
-            return False
-
-        # Step 4: Save tokens
-        self.save_tokens(tokens)
-
-        console.print("\n[green]Successfully authenticated with Basic Memory Cloud![/green]")
-        return True
-
-    def logout(self) -> None:
-        """Remove stored authentication tokens."""
-        if self.token_file.exists():
-            self.token_file.unlink()
-            console.print("[green]Logged out successfully[/green]")
-        else:
-            console.print("[yellow]No stored authentication found[/yellow]")

src/basic_memory/cli/auto_update.py

@@ -1,391 +0,0 @@
-"""Automatic update checks and upgrades for the Basic Memory CLI."""
-
-from __future__ import annotations
-
-import json
-import subprocess
-import sys
-import urllib.error
-import urllib.request
-from dataclasses import dataclass
-from datetime import datetime, timedelta
-from enum import Enum
-
-from loguru import logger
-from packaging.version import InvalidVersion, Version
-from rich.console import Console
-
-import basic_memory
-from basic_memory.config import ConfigManager
-
-PACKAGE_NAME = "basic-memory"
-PYPI_JSON_URL = "https://pypi.org/pypi/basic-memory/json"
-
-PYPI_TIMEOUT_SECONDS = 5
-BREW_OUTDATED_TIMEOUT_SECONDS = 15
-UV_UPGRADE_TIMEOUT_SECONDS = 180
-BREW_UPGRADE_TIMEOUT_SECONDS = 600
-
-
-class InstallSource(str, Enum):
-    """How the running CLI appears to have been installed."""
-
-    HOMEBREW = "homebrew"
-    UV_TOOL = "uv_tool"
-    UVX = "uvx"
-    UNKNOWN = "unknown"
-
-
-class AutoUpdateStatus(str, Enum):
-    """Result classification for update checks and installs."""
-
-    SKIPPED = "skipped"
-    UP_TO_DATE = "up_to_date"
-    UPDATE_AVAILABLE = "update_available"
-    UPDATED = "updated"
-    FAILED = "failed"
-
-
-@dataclass(frozen=True)
-class AutoUpdateResult:
-    """Structured result for update checks/install attempts."""
-
-    status: AutoUpdateStatus
-    source: InstallSource
-    checked: bool
-    update_available: bool
-    updated: bool
-    latest_version: str | None = None
-    message: str | None = None
-    error: str | None = None
-    restart_recommended: bool = False
-
-
-def detect_install_source(executable: str | None = None) -> InstallSource:
-    """Infer installation source from the active interpreter path."""
-    active_executable = executable or sys.executable
-    normalized = active_executable.lower().replace("\\", "/")
-
-    if "cellar/basic-memory" in normalized:
-        return InstallSource.HOMEBREW
-    if "uv/tools/basic-memory" in normalized:
-        return InstallSource.UV_TOOL
-    if "/uv/archive-" in normalized:
-        return InstallSource.UVX
-    return InstallSource.UNKNOWN
-
-
-def _is_interactive_session() -> bool:
-    """Return whether stdin/stdout are interactive terminals."""
-    try:
-        return sys.stdin.isatty() and sys.stdout.isatty()
-    except ValueError:
-        # Trigger: stdin/stdout may be closed during transport teardown.
-        # Why: isatty() raises ValueError on closed descriptors.
-        # Outcome: treat as non-interactive and suppress periodic output.
-        return False
-
-
-def _run_subprocess(
-    command: list[str],
-    *,
-    timeout_seconds: int,
-    silent: bool,
-    capture_output: bool,
-) -> subprocess.CompletedProcess[str]:
-    """Run a subprocess with explicit stdio behavior for protocol safety."""
-    # Trigger: silent operation (MCP/background) with no need for subprocess output.
-    # Why: prevent protocol/terminal pollution from child process output.
-    # Outcome: stdout/stderr are discarded unless explicit capture is requested.
-    use_devnull = silent and not capture_output
-    stdout_target = subprocess.DEVNULL if use_devnull else subprocess.PIPE
-    stderr_target = subprocess.DEVNULL if use_devnull else subprocess.PIPE
-
-    return subprocess.run(
-        command,
-        stdin=subprocess.DEVNULL,
-        stdout=stdout_target,
-        stderr=stderr_target,
-        text=True,
-        timeout=timeout_seconds,
-        check=False,
-    )
-
-
-def _version_from_pypi() -> str:
-    """Fetch the latest published package version from PyPI."""
-    request = urllib.request.Request(
-        PYPI_JSON_URL,
-        headers={"User-Agent": f"basic-memory-cli/{basic_memory.__version__}"},
-    )
-    with urllib.request.urlopen(request, timeout=PYPI_TIMEOUT_SECONDS) as response:
-        payload = json.loads(response.read().decode("utf-8"))
-    latest = payload.get("info", {}).get("version")
-    if not latest:
-        raise RuntimeError("PyPI JSON response did not include info.version")
-    return str(latest)
-
-
-def _check_homebrew_update_available(silent: bool) -> tuple[bool, str | None]:
-    """Check whether Homebrew reports an outdated basic-memory formula."""
-    result = _run_subprocess(
-        ["brew", "outdated", "--quiet", PACKAGE_NAME],
-        timeout_seconds=BREW_OUTDATED_TIMEOUT_SECONDS,
-        silent=silent,
-        capture_output=True,
-    )
-    if result.returncode != 0:
-        stderr = (result.stderr or "").strip()
-        stdout = (result.stdout or "").strip()
-        detail = stderr or stdout or "brew outdated failed"
-        raise RuntimeError(detail)
-
-    is_outdated = bool((result.stdout or "").strip())
-    return is_outdated, None
-
-
-def _check_pypi_update_available() -> tuple[bool, str]:
-    """Compare installed package version with PyPI latest version."""
-    latest = _version_from_pypi()
-    try:
-        current_version = Version(basic_memory.__version__)
-        latest_version = Version(latest)
-    except InvalidVersion as exc:
-        raise RuntimeError(
-            f"Could not compare versions (current={basic_memory.__version__}, latest={latest})"
-        ) from exc
-
-    return latest_version > current_version, latest
-
-
-def _manual_update_hint(source: InstallSource) -> str:
-    """Return manager-appropriate manual update instructions."""
-    if source == InstallSource.UV_TOOL:
-        return "Run `uv tool upgrade basic-memory`."
-    if source == InstallSource.HOMEBREW:
-        return "Run `brew upgrade basic-memory`."
-    return (
-        "Automatic install is not supported for this environment. "
-        "Update with your package manager (for pip: `python3 -m pip install -U basic-memory`)."
-    )
-
-
-def _save_last_checked_timestamp(config_manager: ConfigManager, checked_at: datetime) -> None:
-    """Persist the timestamp for the most recent attempted update check."""
-    config = config_manager.load_config()
-    config.auto_update_last_checked_at = checked_at
-    config_manager.save_config(config)
-
-
-def run_auto_update(
-    *,
-    force: bool = False,
-    check_only: bool = False,
-    silent: bool = False,
-    config_manager: ConfigManager | None = None,
-    now: datetime | None = None,
-    executable: str | None = None,
-) -> AutoUpdateResult:
-    """Run update check/install flow and return a structured result."""
-    manager = config_manager or ConfigManager()
-    config = manager.load_config()
-    source = detect_install_source(executable)
-    checked_at = now or datetime.now()
-
-    if source == InstallSource.UVX:
-        return AutoUpdateResult(
-            status=AutoUpdateStatus.SKIPPED,
-            source=source,
-            checked=False,
-            update_available=False,
-            updated=False,
-            message="uvx runtime detected; updates are managed by uvx cache resolution.",
-        )
-
-    if not force and not config.auto_update:
-        return AutoUpdateResult(
-            status=AutoUpdateStatus.SKIPPED,
-            source=source,
-            checked=False,
-            update_available=False,
-            updated=False,
-            message="Auto-update is disabled in config.",
-        )
-
-    if not force and config.auto_update_last_checked_at is not None:
-        try:
-            elapsed = checked_at - config.auto_update_last_checked_at
-        except TypeError:
-            # Trigger: mixed naive/aware datetimes from manual config edits.
-            # Why: datetime subtraction fails for mixed tz-awareness.
-            # Outcome: ignore the gate once and continue with a forced check path.
-            logger.warning("Auto-update interval gate skipped due to incompatible timestamp format")
-        else:
-            if elapsed < timedelta(seconds=config.update_check_interval):
-                return AutoUpdateResult(
-                    status=AutoUpdateStatus.SKIPPED,
-                    source=source,
-                    checked=False,
-                    update_available=False,
-                    updated=False,
-                    message="Update check interval has not elapsed.",
-                )
-
-    try:
-        # --- Availability check ---
-        latest_version: str | None = None
-        if source == InstallSource.HOMEBREW:
-            update_available, latest_version = _check_homebrew_update_available(silent=silent)
-        else:
-            update_available, latest_version = _check_pypi_update_available()
-
-        if not update_available:
-            return AutoUpdateResult(
-                status=AutoUpdateStatus.UP_TO_DATE,
-                source=source,
-                checked=True,
-                update_available=False,
-                updated=False,
-                latest_version=latest_version,
-                message=f"Basic Memory is up to date ({basic_memory.__version__}).",
-            )
-
-        if check_only:
-            return AutoUpdateResult(
-                status=AutoUpdateStatus.UPDATE_AVAILABLE,
-                source=source,
-                checked=True,
-                update_available=True,
-                updated=False,
-                latest_version=latest_version,
-                message=(
-                    f"Update available (latest: {latest_version or 'unknown'}). "
-                    f"{_manual_update_hint(source)}"
-                ),
-            )
-
-        if source == InstallSource.UNKNOWN:
-            return AutoUpdateResult(
-                status=AutoUpdateStatus.UPDATE_AVAILABLE,
-                source=source,
-                checked=True,
-                update_available=True,
-                updated=False,
-                latest_version=latest_version,
-                message=(
-                    f"Update available (latest: {latest_version or 'unknown'}). "
-                    f"{_manual_update_hint(source)}"
-                ),
-            )
-
-        # --- Automatic install ---
-        command = (
-            ["uv", "tool", "upgrade", PACKAGE_NAME]
-            if source == InstallSource.UV_TOOL
-            else ["brew", "upgrade", PACKAGE_NAME]
-        )
-        timeout = (
-            UV_UPGRADE_TIMEOUT_SECONDS
-            if source == InstallSource.UV_TOOL
-            else BREW_UPGRADE_TIMEOUT_SECONDS
-        )
-
-        install_result = _run_subprocess(
-            command,
-            timeout_seconds=timeout,
-            silent=silent,
-            capture_output=not silent,
-        )
-        if install_result.returncode != 0:
-            stderr = (install_result.stderr or "").strip() if install_result.stderr else ""
-            stdout = (install_result.stdout or "").strip() if install_result.stdout else ""
-            detail = stderr or stdout or "update command failed"
-            return AutoUpdateResult(
-                status=AutoUpdateStatus.FAILED,
-                source=source,
-                checked=True,
-                update_available=True,
-                updated=False,
-                latest_version=latest_version,
-                message="Automatic update failed.",
-                error=detail,
-            )
-
-        return AutoUpdateResult(
-            status=AutoUpdateStatus.UPDATED,
-            source=source,
-            checked=True,
-            update_available=True,
-            updated=True,
-            latest_version=latest_version,
-            message=(
-                "Basic Memory was updated successfully. "
-                "Restart running sessions to use the new version."
-            ),
-            restart_recommended=True,
-        )
-
-    except (
-        RuntimeError,
-        urllib.error.URLError,
-        ValueError,
-        TimeoutError,
-        subprocess.SubprocessError,
-        OSError,
-    ) as exc:
-        logger.warning(f"Auto-update check failed: {exc}")
-        return AutoUpdateResult(
-            status=AutoUpdateStatus.FAILED,
-            source=source,
-            checked=True,
-            update_available=False,
-            updated=False,
-            message="Automatic update check failed.",
-            error=str(exc),
-        )
-    finally:
-        # Trigger: we attempted a check path (including failures).
-        # Why: repeated failing checks on every command create noise and unnecessary network load.
-        # Outcome: next periodic check is gated by update_check_interval.
-        try:
-            _save_last_checked_timestamp(manager, checked_at)
-        except Exception as exc:  # pragma: no cover
-            logger.warning(f"Failed to persist auto-update timestamp: {exc}")
-
-
-def maybe_run_periodic_auto_update(
-    invoked_subcommand: str | None,
-    *,
-    config_manager: ConfigManager | None = None,
-    is_interactive: bool | None = None,
-    console: Console | None = None,
-) -> AutoUpdateResult | None:
-    """Run a periodic auto-update check for interactive CLI sessions."""
-    interactive = _is_interactive_session() if is_interactive is None else is_interactive
-    if not interactive:
-        return None
-    if invoked_subcommand in {None, "mcp", "update"}:
-        return None
-
-    result = run_auto_update(
-        force=False,
-        check_only=False,
-        silent=False,
-        config_manager=config_manager,
-    )
-
-    if result.status in {
-        AutoUpdateStatus.UPDATE_AVAILABLE,
-        AutoUpdateStatus.UPDATED,
-        AutoUpdateStatus.FAILED,
-    }:
-        out = console or Console()
-        if result.status == AutoUpdateStatus.UPDATED:
-            out.print(f"[green]{result.message}[/green]")
-        elif result.status == AutoUpdateStatus.FAILED:
-            error_detail = f" {result.error}" if result.error else ""
-            out.print(f"[yellow]{result.message}{error_detail}[/yellow]")
-        elif result.message:
-            out.print(f"[cyan]{result.message}[/cyan]")
-
-    return result

src/basic_memory/cli/commands/__init__.py

@@ -1,20 +1,13 @@
 """CLI commands for basic-memory."""
 
-from . import status, db, doctor, import_memory_json, mcp, import_claude_conversations
-from . import (
-    import_claude_projects,
-    import_chatgpt,
-    tool,
-    project,
-    format,
-    schema,
-    update,
-)
+from . import auth, status, sync, db, import_memory_json, mcp, import_claude_conversations
+from . import import_claude_projects, import_chatgpt, tool, project
 
 __all__ = [
+    "auth",
     "status",
+    "sync",
     "db",
-    "doctor",
     "import_memory_json",
     "mcp",
     "import_claude_conversations",
@@ -22,7 +15,4 @@ __all__ = [
     "import_chatgpt",
     "tool",
     "project",
-    "format",
-    "schema",
-    "update",
 ]

src/basic_memory/cli/commands/auth.py

@@ -0,0 +1,136 @@
+"""OAuth management commands."""
+
+import typer
+from typing import Optional
+from pydantic import AnyHttpUrl
+
+from basic_memory.cli.app import app
+from basic_memory.mcp.auth_provider import BasicMemoryOAuthProvider
+from mcp.shared.auth import OAuthClientInformationFull
+
+
+auth_app = typer.Typer(help="OAuth client management commands")
+app.add_typer(auth_app, name="auth")
+
+
+@auth_app.command()
+def register_client(
+    client_id: Optional[str] = typer.Option(
+        None, help="Client ID (auto-generated if not provided)"
+    ),
+    client_secret: Optional[str] = typer.Option(
+        None, help="Client secret (auto-generated if not provided)"
+    ),
+    issuer_url: str = typer.Option("http://localhost:8000", help="OAuth issuer URL"),
+):
+    """Register a new OAuth client for Basic Memory MCP server."""
+
+    # Create provider instance
+    provider = BasicMemoryOAuthProvider(issuer_url=issuer_url)
+
+    # Create client info with required redirect_uris
+    client_info = OAuthClientInformationFull(
+        client_id=client_id or "",  # Provider will generate if empty
+        client_secret=client_secret or "",  # Provider will generate if empty
+        redirect_uris=[AnyHttpUrl("http://localhost:8000/callback")],  # Default redirect URI
+        client_name="Basic Memory OAuth Client",
+        grant_types=["authorization_code", "refresh_token"],
+    )
+
+    # Register the client
+    import asyncio
+
+    asyncio.run(provider.register_client(client_info))
+
+    typer.echo("Client registered successfully!")
+    typer.echo(f"Client ID: {client_info.client_id}")
+    typer.echo(f"Client Secret: {client_info.client_secret}")
+    typer.echo("\nSave these credentials securely - the client secret cannot be retrieved later.")
+
+
+@auth_app.command()
+def test_auth(
+    issuer_url: str = typer.Option("http://localhost:8000", help="OAuth issuer URL"),
+):
+    """Test OAuth authentication flow.
+
+    IMPORTANT: Use the same FASTMCP_AUTH_SECRET_KEY environment variable
+    as your MCP server for tokens to validate correctly.
+    """
+
+    import asyncio
+    import secrets
+    from mcp.server.auth.provider import AuthorizationParams
+    from pydantic import AnyHttpUrl
+
+    async def test_flow():
+        # Create provider with same secret key as server
+        provider = BasicMemoryOAuthProvider(issuer_url=issuer_url)
+
+        # Register a test client
+        client_info = OAuthClientInformationFull(
+            client_id=secrets.token_urlsafe(16),
+            client_secret=secrets.token_urlsafe(32),
+            redirect_uris=[AnyHttpUrl("http://localhost:8000/callback")],
+            client_name="Test OAuth Client",
+            grant_types=["authorization_code", "refresh_token"],
+        )
+        await provider.register_client(client_info)
+        typer.echo(f"Registered test client: {client_info.client_id}")
+
+        # Get the client
+        client = await provider.get_client(client_info.client_id)
+        if not client:
+            typer.echo("Error: Client not found after registration", err=True)
+            return
+
+        # Create authorization request
+        auth_params = AuthorizationParams(
+            state="test-state",
+            scopes=["read", "write"],
+            code_challenge="test-challenge",
+            redirect_uri=AnyHttpUrl("http://localhost:8000/callback"),
+            redirect_uri_provided_explicitly=True,
+        )
+
+        # Get authorization URL
+        auth_url = await provider.authorize(client, auth_params)
+        typer.echo(f"Authorization URL: {auth_url}")
+
+        # Extract auth code from URL
+        from urllib.parse import urlparse, parse_qs
+
+        parsed = urlparse(auth_url)
+        params = parse_qs(parsed.query)
+        auth_code = params.get("code", [None])[0]
+
+        if not auth_code:
+            typer.echo("Error: No authorization code in URL", err=True)
+            return
+
+        # Load the authorization code
+        code_obj = await provider.load_authorization_code(client, auth_code)
+        if not code_obj:
+            typer.echo("Error: Invalid authorization code", err=True)
+            return
+
+        # Exchange for tokens
+        token = await provider.exchange_authorization_code(client, code_obj)
+        typer.echo(f"Access token: {token.access_token}")
+        typer.echo(f"Refresh token: {token.refresh_token}")
+        typer.echo(f"Expires in: {token.expires_in} seconds")
+
+        # Validate access token
+        access_token_obj = await provider.load_access_token(token.access_token)
+        if access_token_obj:
+            typer.echo("Access token validated successfully!")
+            typer.echo(f"Client ID: {access_token_obj.client_id}")
+            typer.echo(f"Scopes: {access_token_obj.scopes}")
+        else:
+            typer.echo("Error: Invalid access token", err=True)
+
+    asyncio.run(test_flow())
+
+
+if __name__ == "__main__":
+    auth_app()

src/basic_memory/cli/commands/cloud/__init__.py

@@ -1,19 +0,0 @@
-"""Cloud commands package."""
-
-from basic_memory.cli.app import cloud_app
-
-# Import all commands to register them with typer
-from basic_memory.cli.commands.cloud.core_commands import *  # noqa: F401,F403
-from basic_memory.cli.commands.cloud.api_client import get_authenticated_headers, get_cloud_config  # noqa: F401
-from basic_memory.cli.commands.cloud.upload_command import *  # noqa: F401,F403
-from basic_memory.cli.commands.cloud.project_sync import *  # noqa: F401,F403
-
-# Register snapshot sub-command group
-from basic_memory.cli.commands.cloud.snapshot import snapshot_app
-from basic_memory.cli.commands.cloud.workspace import workspace_app
-
-cloud_app.add_typer(snapshot_app, name="snapshot")
-cloud_app.add_typer(workspace_app, name="workspace")
-
-# Register restore command (directly on cloud_app via decorator)
-from basic_memory.cli.commands.cloud.restore import restore  # noqa: F401, E402

src/basic_memory/cli/commands/cloud/api_client.py

@@ -1,127 +0,0 @@
-"""Cloud API client utilities."""
-
-from collections.abc import AsyncIterator
-from typing import Optional
-from contextlib import asynccontextmanager
-from typing import AsyncContextManager, Callable
-
-import httpx
-import typer
-from rich.console import Console
-
-from basic_memory.cli.auth import CLIAuth
-from basic_memory.config import ConfigManager
-
-console = Console()
-
-HttpClientFactory = Callable[[], AsyncContextManager[httpx.AsyncClient]]
-
-
-class CloudAPIError(Exception):
-    """Exception raised for cloud API errors."""
-
-    def __init__(
-        self, message: str, status_code: Optional[int] = None, detail: Optional[dict] = None
-    ):
-        super().__init__(message)
-        self.status_code = status_code
-        self.detail = detail or {}
-
-
-class SubscriptionRequiredError(CloudAPIError):
-    """Exception raised when user needs an active subscription."""
-
-    def __init__(self, message: str, subscribe_url: str):
-        super().__init__(message, status_code=403, detail={"error": "subscription_required"})
-        self.subscribe_url = subscribe_url
-
-
-def get_cloud_config() -> tuple[str, str, str]:
-    """Get cloud OAuth configuration from config."""
-    config_manager = ConfigManager()
-    config = config_manager.config
-    return config.cloud_client_id, config.cloud_domain, config.cloud_host
-
-
-async def get_authenticated_headers(auth: CLIAuth | None = None) -> dict[str, str]:
-    """
-    Get authentication headers with JWT token.
-    handles jwt refresh if needed.
-    """
-    client_id, domain, _ = get_cloud_config()
-    auth_obj = auth or CLIAuth(client_id=client_id, authkit_domain=domain)
-    token = await auth_obj.get_valid_token()
-    if not token:
-        console.print("[red]Not authenticated. Please run 'bm cloud login' first.[/red]")
-        raise typer.Exit(1)
-
-    return {"Authorization": f"Bearer {token}"}
-
-
-@asynccontextmanager
-async def _default_http_client(timeout: float) -> AsyncIterator[httpx.AsyncClient]:
-    async with httpx.AsyncClient(timeout=timeout) as client:
-        yield client
-
-
-async def make_api_request(
-    method: str,
-    url: str,
-    headers: Optional[dict] = None,
-    json_data: Optional[dict] = None,
-    timeout: float = 30.0,
-    *,
-    auth: CLIAuth | None = None,
-    http_client_factory: HttpClientFactory | None = None,
-) -> httpx.Response:
-    """Make an API request to the cloud service."""
-    headers = headers or {}
-    auth_headers = await get_authenticated_headers(auth=auth)
-    headers.update(auth_headers)
-    # Add debug headers to help with compression issues
-    headers.setdefault("Accept-Encoding", "identity")  # Disable compression for debugging
-
-    client_factory = http_client_factory or (lambda: _default_http_client(timeout))
-    async with client_factory() as client:
-        try:
-            response = await client.request(method=method, url=url, headers=headers, json=json_data)
-            response.raise_for_status()
-            return response
-        except httpx.HTTPError as e:
-            # Check if this is a response error with response details
-            if hasattr(e, "response") and e.response is not None:  # pyright: ignore [reportAttributeAccessIssue]
-                response = e.response  # type: ignore
-
-                # Try to parse error detail from response
-                error_detail = None
-                try:
-                    error_detail = response.json()
-                except Exception:
-                    # If JSON parsing fails, we'll handle it as a generic error
-                    pass
-
-                # Check for subscription_required error (403)
-                if response.status_code == 403 and isinstance(error_detail, dict):
-                    # Handle both FastAPI HTTPException format (nested under "detail")
-                    # and direct format
-                    detail_obj = error_detail.get("detail", error_detail)
-                    if (
-                        isinstance(detail_obj, dict)
-                        and detail_obj.get("error") == "subscription_required"
-                    ):
-                        message = detail_obj.get("message", "Active subscription required")
-                        subscribe_url = detail_obj.get(
-                            "subscribe_url", "https://basicmemory.com/subscribe"
-                        )
-                        raise SubscriptionRequiredError(
-                            message=message, subscribe_url=subscribe_url
-                        ) from e
-
-                # Raise generic CloudAPIError with status code and detail
-                raise CloudAPIError(
-                    f"API request failed: {e}",
-                    status_code=response.status_code,
-                    detail=error_detail if isinstance(error_detail, dict) else {},
-                ) from e
-
-            raise CloudAPIError(f"API request failed: {e}") from e

src/basic_memory/cli/commands/cloud/bisync_commands.py

@@ -1,110 +0,0 @@
-"""Cloud bisync utility functions for Basic Memory CLI."""
-
-from pathlib import Path
-
-from basic_memory.cli.commands.cloud.api_client import make_api_request
-from basic_memory.config import ConfigManager
-from basic_memory.ignore_utils import create_default_bmignore, get_bmignore_path
-from basic_memory.schemas.cloud import MountCredentials, TenantMountInfo
-
-
-class BisyncError(Exception):
-    """Exception raised for bisync-related errors."""
-
-    pass
-
-
-async def get_mount_info() -> TenantMountInfo:
-    """Get current tenant information from cloud API."""
-    try:
-        config_manager = ConfigManager()
-        config = config_manager.config
-        host_url = config.cloud_host.rstrip("/")
-
-        response = await make_api_request(method="GET", url=f"{host_url}/tenant/mount/info")
-
-        return TenantMountInfo.model_validate(response.json())
-    except Exception as e:
-        raise BisyncError(f"Failed to get tenant info: {e}") from e
-
-
-async def generate_mount_credentials(tenant_id: str) -> MountCredentials:
-    """Generate scoped credentials for syncing."""
-    try:
-        config_manager = ConfigManager()
-        config = config_manager.config
-        host_url = config.cloud_host.rstrip("/")
-
-        response = await make_api_request(method="POST", url=f"{host_url}/tenant/mount/credentials")
-
-        return MountCredentials.model_validate(response.json())
-    except Exception as e:
-        raise BisyncError(f"Failed to generate credentials: {e}") from e
-
-
-def convert_bmignore_to_rclone_filters() -> Path:
-    """Convert .bmignore patterns to rclone filter format.
-
-    Reads ~/.basic-memory/.bmignore (gitignore-style) and converts to
-    ~/.basic-memory/.bmignore.rclone (rclone filter format).
-
-    Only regenerates if .bmignore has been modified since last conversion.
-
-    Returns:
-        Path to converted rclone filter file
-    """
-    # Ensure .bmignore exists
-    create_default_bmignore()
-
-    bmignore_path = get_bmignore_path()
-    # Create rclone filter path: ~/.basic-memory/.bmignore -> ~/.basic-memory/.bmignore.rclone
-    rclone_filter_path = bmignore_path.parent / f"{bmignore_path.name}.rclone"
-
-    # Skip regeneration if rclone file is newer than bmignore
-    if rclone_filter_path.exists():
-        bmignore_mtime = bmignore_path.stat().st_mtime
-        rclone_mtime = rclone_filter_path.stat().st_mtime
-        if rclone_mtime >= bmignore_mtime:
-            return rclone_filter_path
-
-    # Read .bmignore patterns
-    patterns = []
-    try:
-        with bmignore_path.open("r", encoding="utf-8") as f:
-            for line in f:
-                line = line.strip()
-                # Keep comments and empty lines
-                if not line or line.startswith("#"):
-                    patterns.append(line)
-                    continue
-
-                # Convert gitignore pattern to rclone filter syntax
-                # gitignore: node_modules  → rclone: - node_modules/**
-                # gitignore: *.pyc        → rclone: - *.pyc
-                if "*" in line:
-                    # Pattern already has wildcard, just add exclude prefix
-                    patterns.append(f"- {line}")
-                else:
-                    # Directory pattern - add /** for recursive exclude
-                    patterns.append(f"- {line}/**")
-
-    except Exception:
-        # If we can't read the file, create a minimal filter
-        patterns = ["# Error reading .bmignore, using minimal filters", "- .git/**"]
-
-    # Write rclone filter file
-    rclone_filter_path.write_text("\n".join(patterns) + "\n")
-
-    return rclone_filter_path
-
-
-def get_bisync_filter_path() -> Path:
-    """Get path to bisync filter file.
-
-    Uses ~/.basic-memory/.bmignore (converted to rclone format).
-    The file is automatically created with default patterns on first use.
-
-    Returns:
-        Path to rclone filter file
-    """
-    return convert_bmignore_to_rclone_filters()

src/basic_memory/cli/commands/cloud/cloud_utils.py

@@ -1,108 +0,0 @@
-"""Shared utilities for cloud operations."""
-
-from basic_memory.cli.commands.cloud.api_client import make_api_request
-from basic_memory.config import ConfigManager
-from basic_memory.schemas.cloud import (
-    CloudProjectList,
-    CloudProjectCreateRequest,
-    CloudProjectCreateResponse,
-)
-from basic_memory.utils import generate_permalink
-
-
-class CloudUtilsError(Exception):
-    """Exception raised for cloud utility errors."""
-
-    pass
-
-
-async def fetch_cloud_projects(
-    *,
-    api_request=make_api_request,
-) -> CloudProjectList:
-    """Fetch list of projects from cloud API.
-
-    Returns:
-        CloudProjectList with projects from cloud
-    """
-    try:
-        config_manager = ConfigManager()
-        config = config_manager.config
-        host_url = config.cloud_host.rstrip("/")
-
-        response = await api_request(method="GET", url=f"{host_url}/proxy/v2/projects/")
-
-        return CloudProjectList.model_validate(response.json())
-    except Exception as e:
-        raise CloudUtilsError(f"Failed to fetch cloud projects: {e}") from e
-
-
-async def create_cloud_project(
-    project_name: str,
-    *,
-    api_request=make_api_request,
-) -> CloudProjectCreateResponse:
-    """Create a new project on cloud.
-
-    Args:
-        project_name: Name of project to create
-
-    Returns:
-        CloudProjectCreateResponse with project details from API
-    """
-    try:
-        config_manager = ConfigManager()
-        config = config_manager.config
-        host_url = config.cloud_host.rstrip("/")
-
-        # Use generate_permalink to ensure consistent naming
-        project_path = generate_permalink(project_name)
-
-        project_data = CloudProjectCreateRequest(
-            name=project_name,
-            path=project_path,
-            set_default=False,
-        )
-
-        response = await api_request(
-            method="POST",
-            url=f"{host_url}/proxy/v2/projects/",
-            headers={"Content-Type": "application/json"},
-            json_data=project_data.model_dump(),
-        )
-
-        return CloudProjectCreateResponse.model_validate(response.json())
-    except Exception as e:
-        raise CloudUtilsError(f"Failed to create cloud project '{project_name}': {e}") from e
-
-
-async def sync_project(project_name: str, force_full: bool = False) -> None:
-    """Trigger sync for a specific project on cloud.
-
-    Args:
-        project_name: Name of project to sync
-        force_full: If True, force a full scan bypassing watermark optimization
-    """
-    try:
-        from basic_memory.cli.commands.command_utils import run_sync
-
-        await run_sync(project=project_name, force_full=force_full)
-    except Exception as e:
-        raise CloudUtilsError(f"Failed to sync project '{project_name}': {e}") from e
-
-
-async def project_exists(project_name: str, *, api_request=make_api_request) -> bool:
-    """Check if a project exists on cloud.
-
-    Args:
-        project_name: Name of project to check
-
-    Returns:
-        True if project exists, False otherwise
-    """
-    try:
-        projects = await fetch_cloud_projects(api_request=api_request)
-        project_names = {p.name for p in projects.projects}
-        return project_name in project_names
-    except Exception:
-        return False