Revert "feat: Add project-prefixed permalinks and memory URL routing (#544)"

This reverts commit 545804f194.

.claude/commands/check-health.md

@@ -1,190 +0,0 @@
-# /project:check-health - Project Health Assessment
-
-Comprehensive health check of the Basic Memory project including code quality, test coverage, dependencies, and documentation.
-
-## Usage
-```
-/project:check-health
-```
-
-## Implementation
-
-You are an expert DevOps engineer for the Basic Memory project. When the user runs `/project:check-health`, execute the following comprehensive assessment:
-
-### Step 1: Git Repository Health
-1. **Repository Status**
-   ```bash
-   git status
-   git log --oneline -5
-   git branch -vv
-   ```
-   - Check working directory status
-   - Verify branch alignment with remote
-   - Check recent commit activity
-
-2. **Branch Analysis**
-   - Verify on main branch
-   - Check if ahead/behind remote
-   - Identify any untracked files
-
-### Step 2: Code Quality Assessment
-1. **Linting and Formatting**
-   ```bash
-   uv run ruff check .
-   uv run pyright
-   ```
-   - Count linting issues by severity
-   - Check type annotation coverage
-   - Verify code formatting compliance
-
-2. **Test Suite Health**
-   ```bash
-   uv run pytest --collect-only -q
-   uv run pytest --co -q | wc -l
-   ```
-   - Count total tests
-   - Check for test discovery issues
-   - Verify test structure integrity
-
-### Step 3: Dependency Analysis
-1. **Dependency Health**
-   ```bash
-   uv tree
-   uv lock --dry-run
-   ```
-   - Check for dependency conflicts
-   - Identify outdated dependencies
-   - Verify lock file consistency
-
-2. **Security Scan**
-   ```bash
-   uv run pip-audit --desc
-   ```
-   - Scan for known vulnerabilities
-   - Check dependency licenses
-   - Identify security advisories
-
-### Step 4: Performance Metrics
-1. **Test Performance**
-   ```bash
-   uv run pytest --durations=10
-   ```
-   - Identify slowest tests
-   - Check overall test execution time
-   - Monitor test suite growth
-
-2. **Build Performance**
-   ```bash
-   time uv run python -c "import basic_memory"
-   ```
-   - Check import time
-   - Validate package installation
-   - Monitor startup performance
-
-### Step 5: Documentation Health
-1. **Documentation Coverage**
-   - Check README.md currency
-   - Verify CLI documentation
-   - Validate MCP tool documentation
-   - Check changelog completeness
-
-2. **API Documentation**
-   - Verify docstring coverage
-   - Check type annotation completeness
-   - Validate example code
-
-### Step 6: Project Metrics
-1. **Code Statistics**
-   ```bash
-   find src -name "*.py" | xargs wc -l
-   find tests -name "*.py" | xargs wc -l
-   ```
-   - Lines of code trends
-   - Test-to-code ratio
-   - File organization metrics
-
-## Health Report Format
-
-Generate comprehensive health dashboard:
-
-```
-🏥 Basic Memory Project Health Report
-
-📊 OVERALL HEALTH: 🟢 EXCELLENT (92/100)
-
-🗂️  GIT REPOSITORY
-✅ Clean working directory
-✅ Up to date with origin/main
-✅ Recent commit activity (5 commits this week)
-
-🔍 CODE QUALITY  
-✅ Linting: 0 errors, 2 warnings
-✅ Type checking: 100% coverage
-✅ Formatting: Compliant
-⚠️  Complex functions: 3 need refactoring
-
-🧪 TEST SUITE
-✅ Total tests: 744 
-✅ Test discovery: All tests found
-✅ Coverage: 98.2%
-⚡ Performance: 45.2s (good)
-
-📦 DEPENDENCIES
-✅ Dependencies: Up to date
-✅ Security: No vulnerabilities
-✅ Conflicts: None detected
-⚠️  Outdated: 2 minor updates available
-
-📖 DOCUMENTATION
-✅ README: Current
-✅ API docs: 95% coverage
-⚠️  CLI reference: Needs update
-✅ Changelog: Complete
-
-📈 METRICS
-├── Source code: 15,432 lines
-├── Test code: 8,967 lines  
-├── Test ratio: 58% (excellent)
-└── Complexity: Low (maintainable)
-
-🎯 RECOMMENDATIONS:
-1. Update CLI documentation
-2. Refactor 3 complex functions
-3. Update minor dependencies
-4. Consider splitting large test files
-
-🏆 PROJECT STATUS: Ready for v0.13.0 release!
-```
-
-## Health Scoring
-
-### Excellent (90-100)
-- All quality gates pass
-- High test coverage (>95%)
-- No security issues
-- Documentation current
-
-### Good (75-89)
-- Minor issues present
-- Good test coverage (>90%)
-- No critical security issues
-- Most documentation current
-
-### Needs Attention (60-74)
-- Several quality issues
-- Adequate test coverage (>80%)
-- Minor security concerns
-- Documentation gaps
-
-### Critical (<60)
-- Major quality problems
-- Low test coverage (<80%)
-- Security vulnerabilities
-- Significant documentation issues
-
-## Context
-- Provides comprehensive project overview
-- Identifies potential issues before they become problems
-- Tracks project health trends over time
-- Helps prioritize maintenance tasks
-- Supports release readiness decisions

.claude/commands/commands.md

@@ -1,62 +0,0 @@
-# Basic Memory Custom Commands
-
-This directory contains custom Claude Code slash commands for the Basic Memory project.
-
-## Available Commands
-
-### Release Management (`/project:release:*`)
-- `/project:release:beta` - Create beta releases with automated quality checks
-- `/project:release:release` - Create stable releases with comprehensive validation  
-- `/project:release:release-check` - Pre-flight validation without making changes
-- `/project:release:changelog` - Generate changelog entries from commits
-
-### Development (`/project:*`)
-- `/project:test-coverage` - Run tests with detailed coverage analysis
-- `/project:test-live` - Live testing suite using real Basic Memory installation
-- `/project:lint-fix` - Run comprehensive linting with auto-fix
-- `/project:check-health` - Comprehensive project health assessment
-
-## Command Structure
-
-Commands are organized by functionality:
-```
-.claude/commands/
-├── release/          # Release management commands
-│   ├── beta.md       # /project:release:beta
-│   ├── release.md    # /project:release:release
-│   ├── release-check.md # /project:release:release-check
-│   └── changelog.md  # /project:release:changelog
-├── test-coverage.md  # /project:test-coverage
-├── test-live.md      # /project:test-live
-├── lint-fix.md       # /project:lint-fix
-├── check-health.md   # /project:check-health
-└── commands.md       # This overview file
-```
-
-## Usage
-
-Commands are invoked using the `/project:` prefix:
-- `/project:release:beta v0.13.0b4`
-- `/project:test-coverage mcp`
-- `/project:test-live core`
-- `/project:release:release-check`
-- `/project:check-health`
-
-## Implementation
-
-Each command is implemented as a Markdown file containing structured prompts that:
-1. Validate preconditions
-2. Execute steps in the correct order
-3. Handle errors gracefully  
-4. Provide clear status updates
-5. Return actionable results
-
-## Tooling Integration
-
-Commands leverage existing project tooling:
-- `just check` - Quality checks
-- `just test` - Test suite
-- `just update-deps` - Dependency updates  
-- `uv` - Package management
-- `git` - Version control
-- GitHub Actions - CI/CD pipeline

.claude/commands/lint-fix.md

@@ -1,145 +0,0 @@
-# /project:lint-fix - Comprehensive Code Quality Fix
-
-Run comprehensive linting and auto-fix code quality issues across the codebase.
-
-## Usage
-```
-/project:lint-fix
-```
-
-## Implementation
-
-You are an expert code quality engineer for the Basic Memory project. When the user runs `/project:lint-fix`, execute the following steps:
-
-### Step 1: Pre-flight Check
-1. **Verify Clean Working Directory**
-   ```bash
-   git status --porcelain
-   ```
-   - Check for uncommitted changes
-   - Warn if working directory is not clean
-   - Suggest stashing changes if needed
-
-### Step 2: Import Organization
-1. **Fix Import Order and Cleanup**
-   ```bash
-   uv run ruff check --select I --fix .
-   ```
-   - Sort imports by category (standard, third-party, local)
-   - Remove unused imports
-   - Fix import spacing and organization
-
-### Step 3: Code Formatting
-1. **Apply Consistent Formatting**
-   ```bash
-   uv run ruff format .
-   ```
-   - Format code according to project style
-   - Fix line length issues (100 chars max)
-   - Standardize quotes and spacing
-
-### Step 4: Linting with Auto-fix
-1. **Fix Linting Issues**
-   ```bash
-   uv run ruff check --fix .
-   ```
-   - Auto-fix safe linting issues
-   - Report any remaining manual fixes needed
-   - Focus on code quality and best practices
-
-### Step 5: Type Checking
-1. **Validate Type Annotations**
-   ```bash
-   uv run pyright
-   ```
-   - Check for type errors
-   - Report any missing type annotations
-   - Validate type compatibility
-
-### Step 6: Report Generation
-Generate comprehensive quality report:
-
-```
-🔧 Code Quality Fix Report
-
-✅ FIXES APPLIED:
-├── Import organization: 12 files updated
-├── Code formatting: 8 files reformatted  
-├── Auto-fixable lint issues: 23 issues resolved
-└── Total files processed: 156
-
-⚠️  MANUAL ATTENTION NEEDED:
-├── Type annotations missing in entity_service.py:45
-├── Complex function needs refactoring in sync_service.py:123
-└── Unused variable in test_utils.py:67
-
-🎯 QUALITY SCORE: 96.2% (excellent)
-
-📁 Run `git diff` to review all changes
-```
-
-## Error Handling
-
-### Common Issues
-- **Merge Conflicts**: Provide resolution guidance
-- **Syntax Errors**: Point to specific files and lines
-- **Type Errors**: Suggest specific fixes
-- **Import Errors**: Check for missing dependencies
-
-### Recovery Steps
-- If auto-fixes introduce issues, provide rollback instructions
-- If type checking fails, suggest incremental fixes
-- If tests break, provide debugging guidance
-
-## Quality Gates
-
-### Must Pass
-- [ ] All auto-fixable lint issues resolved
-- [ ] Code formatting consistent
-- [ ] No syntax errors
-- [ ] Import organization clean
-
-### Should Pass (Warnings)
-- [ ] No type checking errors
-- [ ] No complex function warnings
-- [ ] No unused variables/imports
-- [ ] Consistent naming conventions
-
-## Output Examples
-
-### Successful Fix
-```
-🎉 CODE QUALITY IMPROVED!
-
-✅ All auto-fixes applied successfully
-📏 Code formatting: 100% compliant
-🔍 Linting: No issues found
-🏷️  Type checking: All passed
-
-Ready for commit! Use:
-git add -A && git commit -m "style: fix code quality issues"
-```
-
-### Issues Requiring Attention
-```
-⚠️  PARTIAL SUCCESS - MANUAL FIXES NEEDED
-
-✅ Auto-fixes applied: 45 issues
-❌ Manual fixes needed: 3 issues
-
-Priority fixes:
-1. Fix type annotation in services/entity_service.py:142
-2. Simplify complex function in sync/sync_service.py:67
-3. Remove unused import in tests/conftest.py:12
-
-Run these commands:
-# Fix specific file
-uv run pyright src/basic_memory/services/entity_service.py
-```
-
-## Context
-- Uses ruff for fast Python linting and formatting
-- Uses pyright for type checking
-- Follows project code style guidelines (100 char line length)
-- Maintains backward compatibility
-- Integrates with existing pre-commit hooks

.claude/commands/release/beta.md

@@ -1,69 +1,95 @@
 # /beta - Create Beta Release
 
-Create a new beta release for the current version with automated quality checks and tagging.
+Create a new beta release using the automated justfile target with quality checks and tagging.
 
 ## Usage
 ```
-/beta [version]
+/beta <version>
 ```
 
 **Parameters:**
-- `version` (optional): Beta version like `v0.13.0b4`. If not provided, auto-increments from latest beta tag.
+- `version` (required): Beta version like `v0.13.2b1` or `v0.13.2rc1`
 
 ## Implementation
 
 You are an expert release manager for the Basic Memory project. When the user runs `/beta`, execute the following steps:
 
-### Step 1: Pre-flight Checks
-1. Check current git status for uncommitted changes
-2. Verify we're on the `main` branch
-3. Get the latest beta tag to determine next version if not provided
+### Step 1: Pre-flight Validation
+1. Verify version format matches `v\d+\.\d+\.\d+(b\d+|rc\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
 
-### Step 2: Quality Assurance
-1. Run `just check` to ensure code quality
-2. If any checks fail, report issues and stop
-3. Run `just update-deps` to ensure latest dependencies
-4. Commit any dependency updates with proper message
+### Step 2: Use Justfile Automation
+Execute the automated beta release process:
+```bash
+just beta <version>
+```
 
-### Step 3: Version Determination
-If version not provided:
-1. Get latest git tags with `git tag -l "v*b*" --sort=-version:refname | head -1`
-2. Auto-increment beta number (e.g., `v0.13.0b2` → `v0.13.0b3`)
-3. Confirm version with user before proceeding
+The justfile target handles:
+- ✅ Beta version format validation (supports b1, b2, rc1, etc.)
+- ✅ Git status and branch checks
+- ✅ Quality checks (`just check` - lint, format, type-check, tests)
+- ✅ Version update in `src/basic_memory/__init__.py`
+- ✅ Automatic commit with proper message
+- ✅ Tag creation and pushing to GitHub
+- ✅ Beta release workflow trigger
 
-### Step 4: Release Creation
-1. Commit any remaining changes
-2. Push to main: `git push origin main`
-3. Create tag: `git tag {version}`
-4. Push tag: `git push origin {version}`
-
-### Step 5: Monitor Release
+### Step 3: Monitor Beta Release
 1. Check GitHub Actions workflow starts successfully
-2. Provide installation instructions for beta
-3. Report status and next steps
+2. Monitor workflow at: https://github.com/basicmachines-co/basic-memory/actions
+3. Verify PyPI pre-release publication
+4. Test beta installation: `uv tool install basic-memory --pre`
+
+### Step 4: Beta Testing Instructions
+Provide users with beta testing instructions:
+
+```bash
+# Install/upgrade to beta
+uv tool install basic-memory --pre
+
+# Or upgrade existing installation
+uv tool upgrade basic-memory --prerelease=allow
+```
+
+## Version Guidelines
+- **First beta**: `v0.13.2b1` 
+- **Subsequent betas**: `v0.13.2b2`, `v0.13.2b3`, etc.
+- **Release candidates**: `v0.13.2rc1`, `v0.13.2rc2`, etc.
+- **Final release**: `v0.13.2` (use `/release` command)
 
 ## Error Handling
-- If quality checks fail, provide specific fix instructions
-- If git operations fail, provide manual recovery steps  
-- If GitHub Actions fail, provide debugging guidance
+- If `just beta` fails, examine the error output for specific issues
+- If quality checks fail, fix issues and retry
+- If version format is invalid, correct and retry
+- If tag already exists, increment version number
 
 ## Success Output
 ```
-✅ Beta Release v0.13.0b4 Created Successfully!
+✅ Beta Release v0.13.2b1 Created Successfully!
 
-🏷️  Tag: v0.13.0b4
+🏷️  Tag: v0.13.2b1
 🚀 GitHub Actions: Running
-📦 PyPI: Will be available in ~5 minutes
+📦 PyPI: Will be available in ~5 minutes as pre-release
 
-Install with:
-uv tool upgrade basic-memory --prerelease=allow
+Install/test with:
+uv tool install basic-memory --pre
 
 Monitor release: https://github.com/basicmachines-co/basic-memory/actions
 ```
 
+## Beta Testing Workflow
+1. **Create beta**: Use `/beta v0.13.2b1`
+2. **Test features**: Install and validate new functionality
+3. **Fix issues**: Address bugs found during testing
+4. **Iterate**: Create `v0.13.2b2` if needed
+5. **Release candidate**: Create `v0.13.2rc1` when stable
+6. **Final release**: Use `/release v0.13.2` when ready
+
 ## Context
-- Use the existing justfile targets (`just check`, `just update-deps`)
-- Follow semantic versioning for beta releases
-- Maintain release notes in CHANGELOG.md
-- Use conventional commit messages
-- Leverage uv-dynamic-versioning for version management
+- Beta releases are pre-releases for testing new features
+- Automatically published to PyPI with pre-release flag
+- Uses the automated justfile target for consistency
+- Version is automatically updated in `__init__.py`
+- Ideal for validating changes before stable release
+- Supports both beta (b1, b2) and release candidate (rc1, rc2) versions

.claude/commands/release/changelog.md

@@ -8,7 +8,7 @@ Analyze commits and generate formatted changelog entry for a version.
 ```
 
 **Parameters:**
-- `version` (required): Version like `v0.13.0` or `v0.13.0b4`
+- `version` (required): Version like `v0.14.0` or `v0.14.0b1`
 - `type` (optional): `beta`, `rc`, or `stable` (default: `stable`)
 
 ## Implementation
@@ -59,8 +59,9 @@ You are an expert technical writer for the Basic Memory project. When the user r
 ### Step 3: Generate Changelog Entry
 Create formatted entry following existing CHANGELOG.md style:
 
+Example:
 ```markdown
-## v0.13.0 (2025-06-03)
+## <version> (<date>)
 
 ### Features
 
@@ -128,6 +129,8 @@ Create formatted entry following existing CHANGELOG.md style:
 ## Output Format
 
 ### For Beta Releases
+
+Example: 
 ```markdown
 ## v0.13.0b4 (2025-06-03)
 

.claude/commands/release/release.md

@@ -1,6 +1,6 @@
 # /release - Create Stable Release
 
-Create a stable release from the current main branch with comprehensive validation.
+Create a stable release using the automated justfile target with comprehensive validation.
 
 ## Usage
 ```
@@ -8,77 +8,192 @@ Create a stable release from the current main branch with comprehensive validati
 ```
 
 **Parameters:**
-- `version` (required): Release version like `v0.13.0`
+- `version` (required): Release version like `v0.13.2`
 
 ## Implementation
 
 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
-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
 
-### Step 2: Comprehensive Quality Checks
-1. Run `just check` (lint, format, type-check, full test suite)
-2. Verify test coverage meets minimum requirements (95%+)
-3. Check that CHANGELOG.md contains entry for this version
-4. Validate all high-priority issues are closed
+#### 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
 
-### Step 3: Release Preparation
-1. Update any version references if needed
-2. Commit any final changes with message: `chore: prepare for ${version} release`
-3. Push to main: `git push origin main`
+#### 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
 
-### Step 4: Release Creation
-1. Create annotated tag: `git tag -a ${version} -m "Release ${version}"`
-2. Push tag: `git push origin ${version}`
-3. Monitor GitHub Actions for release automation
+#### Documentation Validation
+1. **Changelog Check**
+   - CHANGELOG.md contains entry for target version
+   - Entry includes all major features and fixes
+   - Breaking changes are documented
 
-### Step 5: Post-Release Validation
-1. Verify GitHub release is created automatically
-2. Check PyPI publication
-3. Validate release assets
-4. Test installation: `uv tool install basic-memory`
+### Step 2: Use Justfile Automation
+Execute the automated release process:
+```bash
+just release <version>
+```
 
-### Step 6: Documentation Update
-1. Update any post-release documentation
-2. Create follow-up tasks if needed
+The justfile target handles:
+- ✅ Version format validation
+- ✅ Git status and branch checks
+- ✅ Quality checks (`just check` - lint, format, type-check, tests)
+- ✅ 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)
+
+### 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
+
+### 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
 
 ## Pre-conditions Check
 Before starting, verify:
 - [ ] All beta testing is complete
 - [ ] Critical bugs are fixed
 - [ ] Breaking changes are documented
-- [ ] CHANGELOG.md is updated
+- [ ] CHANGELOG.md is updated (if needed)
 - [ ] Version number follows semantic versioning
 
 ## Error Handling
-- If any quality check fails, stop and provide fix instructions
-- If changelog entry missing, prompt to create one
-- If tests fail, provide debugging guidance
-- If GitHub Actions fail, provide manual release steps
+- If `just release` fails, examine the error output for specific issues
+- If quality checks fail, fix issues and retry
+- If changelog entry missing, update CHANGELOG.md and commit before retrying
+- If GitHub Actions fail, check workflow logs for debugging
 
 ## Success Output
 ```
-🎉 Stable Release v0.13.0 Created Successfully!
+🎉 Stable Release v0.13.2 Created Successfully!
 
-🏷️  Tag: v0.13.0
-📋 GitHub Release: https://github.com/basicmachines-co/basic-memory/releases/tag/v0.13.0
-📦 PyPI: https://pypi.org/project/basic-memory/0.13.0/
+🏷️  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:
-uv tool install basic-memory
+Install with pip/uv:
+  uv tool install basic-memory
+
+Install with Homebrew:
+  brew install basicmachines-co/basic-memory/basic-memory
 
 Users can now upgrade:
-uv tool upgrade basic-memory
+  uv tool upgrade basic-memory
+  brew upgrade basic-memory
 ```
 
 ## Context
 - This creates production releases used by end users
 - Must pass all quality gates before proceeding
-- Follows the release workflow documented in CLAUDE.md
-- Uses uv-dynamic-versioning for automatic version management
-- Triggers automated GitHub release with changelog
+- Uses the automated justfile target for consistency
+- Version is automatically updated in `__init__.py` and `server.json`
+- 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)

.claude/commands/spec.md

@@ -0,0 +1,51 @@
+---
+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-coverage.md

@@ -1,131 +0,0 @@
-# /test-coverage - Run Tests with Coverage Analysis
-
-Execute test suite with comprehensive coverage reporting and analysis.
-
-## Usage
-```
-/test-coverage [pattern]
-```
-
-**Parameters:**
-- `pattern` (optional): Test pattern to run specific tests (e.g., `test_mcp`, `*integration*`)
-
-## Implementation
-
-You are an expert QA engineer for the Basic Memory project. When the user runs `/test-coverage`, execute the following steps:
-
-### Step 1: Test Execution
-1. **Run Tests with Coverage**
-   ```bash
-   # Full test suite
-   uv run pytest --cov=basic_memory --cov-report=html --cov-report=term -v
-   
-   # Or with pattern if provided
-   uv run pytest tests/*{pattern}* --cov=basic_memory --cov-report=html --cov-report=term -v
-   ```
-
-2. **Generate Coverage Reports**
-   - Terminal summary with percentages
-   - HTML report for detailed analysis
-   - Identify files below coverage threshold
-
-### Step 2: Coverage Analysis
-1. **Summary Statistics**
-   - Overall coverage percentage
-   - Number of files with 100% coverage
-   - Files below 95% threshold
-   - Total lines covered/missed
-
-2. **Detailed Breakdown**
-   - Coverage by module/package
-   - Identify untested code paths
-   - Find missing edge case tests
-
-### Step 3: Report Generation
-Generate comprehensive coverage report:
-
-```
-🧪 Test Coverage Report
-
-📊 OVERALL COVERAGE: 98.2% (target: 95%+)
-
-✅ EXCELLENT COVERAGE (>95%):
-├── basic_memory/mcp/: 99.1%
-├── basic_memory/services/: 98.8%
-├── basic_memory/repository/: 97.9%
-└── basic_memory/api/: 96.2%
-
-⚠️  NEEDS ATTENTION (<95%):
-├── basic_memory/sync/: 94.1% (missing 12 lines)
-└── basic_memory/importers/: 91.8% (missing 23 lines)
-
-🎯 SPECIFIC GAPS:
-├── sync_service.py:142-145 (error handling)
-├── importer_base.py:67-70 (edge case)
-└── file_utils.py:89 (exception path)
-
-📁 HTML Report: htmlcov/index.html
-🚀 Run `open htmlcov/index.html` to view detailed report
-```
-
-### Step 4: Actionable Recommendations
-1. **Coverage Improvements**
-   - Suggest specific tests to add
-   - Identify edge cases to cover
-   - Recommend integration tests
-
-2. **Quality Insights**
-   - Highlight well-tested modules
-   - Point out testing patterns to follow
-   - Suggest refactoring for testability
-
-## Advanced Analysis
-
-### Performance Metrics
-- Test execution time by module
-- Slowest tests identification
-- Coverage collection overhead
-
-### Integration Coverage
-- MCP tool integration tests
-- API endpoint coverage
-- Database operation coverage
-- File system operation coverage
-
-## Output Examples
-
-### Full Coverage Success
-```
-🎉 EXCELLENT COVERAGE!
-
-📊 Coverage: 98.7% (744 tests passed)
-✅ All modules above 95% threshold
-🏆 23 files with 100% coverage
-⚡ Tests completed in 45.2s
-
-Ready for release! 🚀
-```
-
-### Coverage Issues Found
-```
-⚠️  COVERAGE GAPS DETECTED
-
-📊 Coverage: 92.1% (below 95% target)
-❌ 3 modules need attention
-🔍 43 uncovered lines found
-
-Priority fixes:
-1. Add tests for error handling in sync_service.py
-2. Cover edge cases in importer_base.py
-3. Test exception paths in file_utils.py
-
-Run specific tests:
-uv run pytest tests/sync/ -v
-```
-
-## Context
-- Uses pytest with coverage plugin
-- Generates both terminal and HTML reports
-- Focuses on actionable improvement suggestions
-- Integrates with existing test infrastructure
-- Helps maintain high code quality standards

.claude/commands/test-live.md

@@ -1,6 +1,7 @@
 # /project:test-live - Live Basic Memory Testing Suite
 
-Execute comprehensive real-world testing of Basic Memory using the installed version, following the methodology in TESTING.md. All test results are recorded as notes in a dedicated test project.
+Execute comprehensive real-world testing of Basic Memory using the installed version. 
+All test results are recorded as notes in a dedicated test project.
 
 ## Usage
 ```
@@ -8,12 +9,81 @@ Execute comprehensive real-world testing of Basic Memory using the installed ver
 ```
 
 **Parameters:**
-- `phase` (optional): Specific test phase to run (`core`, `features`, `edge`, `workflows`, `stress`, or `all`)
+- `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)
+- `features` - Core + important workflows (Tier 1 + Tier 2)
+- `all` - Comprehensive testing of all tools and scenarios
 
 ## Implementation
 
 You are an expert QA engineer conducting live testing of Basic Memory. 
-When the user runs `/project:test-live`, execute comprehensive testing following the TESTING.md methodology:
+When the user runs `/project:test-live`, execute comprehensive test plan:
+
+## Tool Testing Priority
+
+### **Tier 1: Critical Core (Always Test)**
+1. **write_note** - Foundation of all knowledge creation
+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
+
+### **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
+
+### **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
+
+### **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
 
 ### Pre-Test Setup
 
@@ -22,7 +92,13 @@ When the user runs `/project:test-live`, execute comprehensive testing following
    - Check version and confirm it's the expected release
    - Test MCP connection and tool availability
 
-2. **Test Project Creation**
+2. **Recent Changes Analysis** (if phase includes 'recent' or 'all')
+   - Run `git log --oneline -20` to examine recent commits
+   - Identify new features, bug fixes, and enhancements
+   - Generate targeted test scenarios for recent changes
+   - Prioritize regression testing for recently fixed issues
+
+3. **Test Project Creation**
 
 Run the bash `date` command to get the current date/time. 
 
@@ -32,85 +108,151 @@ Run the bash `date` command to get the current date/time.
    Purpose: Record all test observations and results
    ```
 
-Make sure to switch to the newly created project with the `switch_project()` tool.
+Make sure to use the newly created project for all subsequent test operations by specifying it in the `project` parameter of each tool call.
 
-3. **Baseline Documentation**
+4. **Baseline Documentation**
    Create initial test session note with:
    - Test environment details
    - Version being tested
+   - Recent changes identified (if applicable)
    - Test objectives and scope
    - Start timestamp
 
-### Phase 1: Core Functionality Validation
+### Phase 0: Recent Changes Validation (if 'recent' or 'all' phase)
 
-Test all fundamental MCP tools systematically:
+Based on recent commit analysis, create targeted test scenarios:
 
-**write_note Tests:**
-- Basic note creation with various content types
-- Frontmatter handling (tags, custom fields)
-- Special characters in titles and content
-- Unicode and emoji support
-- Empty notes and minimal content
+**Recent Changes Test Protocol:**
+1. **Feature Addition Tests** - For each new feature identified:
+   - Test basic functionality
+   - Test integration with existing tools
+   - Verify documentation accuracy
+   - Test edge cases and error handling
 
-**read_note Tests:**
-- Read by title, permalink, memory:// URLs
-- Non-existent notes (error handling)
-- Notes with complex formatting
-- Performance with large notes
+2. **Bug Fix Regression Tests** - For each recent fix:
+   - Recreate the original problem scenario
+   - Verify the fix works as expected
+   - Test related functionality isn't broken
+   - Document the verification in test notes
 
-**view_note Tests:**
-- View notes as formatted artifacts (Claude Desktop)
-- Title extraction from frontmatter and headings
-- Unicode and emoji content in artifacts
-- Error handling for non-existent notes
-- Artifact display quality and readability
+3. **Performance/Enhancement Validation** - For optimizations:
+   - Establish baseline timing
+   - Compare with expected improvements
+   - Test under various load conditions
+   - Document performance observations
 
-**search_notes Tests:**
-- Simple text queries
-- Tag-based searches 
-- Boolean operators and complex queries
-- Empty/no results scenarios
-- Performance with growing knowledge base
+**Example Recent Changes (Update based on actual git log):**
+- Watch Service Restart (#156): Test project creation → file modification → automatic restart
+- Cross-Project Moves (#161): Test move_note with cross-project detection
+- Docker Environment Support (#174): Test BASIC_MEMORY_HOME behavior
+- MCP Server Logging (#164): Verify log level configurations
 
-**Recent Activity Tests:**
-- Various timeframes ("today", "1 week", "1d")
-- Type filtering (if available)
-- Empty project scenarios
-- Performance with many recent changes
+### Phase 1: Core Functionality Validation (Tier 1 Tools)
 
-**Context Building Tests:**
-- Different depth levels (1, 2, 3+)
-- Various timeframes
-- Relation traversal accuracy
-- Performance with complex graphs
+Test essential MCP tools that form the foundation of Basic Memory:
 
-### Phase 2: v0.13.0 Feature Deep Dive
+**1. write_note Tests (Critical):**
+- ✅ Basic note creation with frontmatter
+- ✅ Special characters and Unicode in titles
+- ✅ Various content types (lists, headings, code blocks)
+- ✅ Empty notes and minimal content edge cases
+- ⚠️ Error handling for invalid parameters
 
-**Project Management:**
-- Create multiple projects dynamically
-- Switch between projects mid-conversation  
-- Cross-project operations
-- Project discovery and status
-- Default project behavior
-- Invalid project handling
+**2. read_note Tests (Critical):**
+- ✅ Read by title, permalink, memory:// URLs
+- ✅ Non-existent notes (error handling)
+- ✅ Notes with complex markdown formatting
+- ⚠️ Performance with large notes (>10MB)
 
-**Advanced Note Editing:**
-- `edit_note` with append operations
-- Prepend operations
-- Find/replace with validation
-- Section replacement under headers
-- Error scenarios (invalid operations)
-- Frontmatter preservation
+**3. search_notes Tests (Critical):**
+- ✅ Simple text queries across content
+- ✅ Tag-based searches with multiple tags
+- ✅ Boolean operators (AND, OR, NOT)
+- ✅ Empty/no results scenarios
+- ⚠️ Performance with 100+ notes
 
-**File Management:**
-- `move_note` within same project
-- Move between projects
-- Automatic folder creation
-- Special characters in paths
-- Database consistency validation
-- Search index updates after moves
+**4. edit_note Tests (Critical):**
+- ✅ Append operations preserving frontmatter
+- ✅ Prepend operations
+- ✅ Find/replace with validation
+- ✅ Section replacement under headers
+- ⚠️ Error scenarios (invalid operations)
 
-### Phase 3: Edge Case Exploration
+**5. list_memory_projects Tests (Critical):**
+- ✅ Display all projects with clear session guidance
+- ✅ Project discovery workflow prompts
+- ✅ Removal of CLI-specific references
+- ✅ Empty project list handling
+- ✅ Single project constraint mode display
+
+**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
+
+### Phase 2: Important Workflows (Tier 2 Tools)
+
+**7. 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):**
+- ✅ Create projects dynamically
+- ✅ Set default during creation
+- ✅ Path validation and creation
+- ⚠️ Invalid paths and names
+- ✅ Integration with existing projects
+
+**9. 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):**
+- ✅ Background operation monitoring
+- ✅ File synchronization status
+- ✅ Project sync state reporting
+- ⚠️ Error state handling
+
+### Phase 3: Enhanced Functionality (Tier 3 Tools)
+
+**11. 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):**
+- ✅ Raw file content access
+- ✅ Binary file handling
+- ✅ Image file reading
+- ⚠️ Large file performance
+
+**13. delete_note Tests (Enhanced):**
+- ✅ Single note deletion
+- ✅ Database consistency after deletion
+- ⚠️ Non-existent note handling
+- ✅ Confirmation of successful deletion
+
+**14. list_directory Tests (Enhanced):**
+- ✅ Directory content listing
+- ✅ Depth control and filtering
+- ✅ File name globbing
+- ⚠️ Empty directory handling
+
+**15. delete_project Tests (Enhanced):**
+- ✅ Project removal from config
+- ✅ Database cleanup
+- ⚠️ Default project protection
+- ⚠️ Non-existent project handling
+
+### Phase 4: Edge Case Exploration
 
 **Boundary Testing:**
 - Very long titles and content (stress limits)
@@ -134,7 +276,7 @@ Test all fundamental MCP tools systematically:
 - Rapid successive operations
 - Memory usage monitoring
 
-### Phase 4: Real-World Workflow Scenarios
+### Phase 5: Real-World Workflow Scenarios
 
 **Meeting Notes Pipeline:**
 1. Create meeting notes with action items
@@ -154,7 +296,7 @@ Test all fundamental MCP tools systematically:
 1. Technical documentation project
 2. Personal recipe collection project
 3. Learning/course notes project
-4. Switch contexts during conversation
+4. Specify different projects for different operations
 5. Cross-reference related concepts
 
 **Content Evolution:**
@@ -164,7 +306,35 @@ Test all fundamental MCP tools systematically:
 4. Update content with edit operations
 5. Validate knowledge graph integrity
 
-### Phase 5: Creative Stress Testing
+### Phase 6: Specialized Tools Testing (Tier 4)
+
+**16. canvas Tests (Specialized):**
+- ✅ JSON Canvas generation
+- ✅ Node and edge creation
+- ✅ Obsidian compatibility
+- ⚠️ Complex graph handling
+
+**17. MCP Prompts Tests (Specialized):**
+- ✅ ai_assistant_guide output
+- ✅ continue_conversation functionality
+- ✅ Formatted search results
+- ✅ Enhanced activity reports
+
+### Phase 7: Integration & File Watching Tests
+
+**File System Integration:**
+- ✅ Watch service behavior with file changes
+- ✅ Project creation → watch restart (#156)
+- ✅ Multi-project synchronization
+- ⚠️ MCP→API→DB→File stack validation
+
+**Real Integration Testing:**
+- ✅ End-to-end file watching vs manual operations
+- ✅ Cross-session persistence
+- ✅ Database consistency across operations
+- ⚠️ Performance under real file system changes
+
+### Phase 8: Creative Stress Testing
 
 **Creative Exploration:**
 - Rapid project creation/switching patterns
@@ -180,6 +350,26 @@ Test all fundamental MCP tools systematically:
 - Complex boolean search expressions
 - Resource constraint testing
 
+## Test Execution Guidelines
+
+### Quick Testing (core/features phases)
+- Focus on Tier 1 tools (core) or Tier 1+2 (features)
+- Test essential functionality and common edge cases
+- Record critical issues immediately
+- Complete in 15-20 minutes
+
+### Comprehensive Testing (all phase)
+- Cover all tiers systematically
+- Include specialized tools and stress testing
+- Document performance baselines
+- Complete in 45-60 minutes
+
+### Recent Changes Focus (recent phase)
+- Analyze git log for recent commits
+- Generate targeted test scenarios
+- Focus on regression testing for fixes
+- Validate new features thoroughly
+
 ## Test Observation Format
 
 Record ALL observations immediately as Basic Memory notes:
@@ -202,14 +392,14 @@ permalink: test-session-[phase]-[timestamp]
 ## Test Results
 
 ### ✅ Successful Operations
-- [timestamp] write_note: Created note with emoji title 📝 #functionality
-- [timestamp] search_notes: Boolean query returned 23 results in 0.4s #performance  
-- [timestamp] edit_note: Append operation preserved frontmatter #reliability
+- [timestamp] ✅ write_note: Created note with emoji title 📝 #tier1 #functionality
+- [timestamp] ✅ search_notes: Boolean query returned 23 results in 0.4s #tier1 #performance  
+- [timestamp] ✅ edit_note: Append operation preserved frontmatter #tier1 #reliability
 
 ### ⚠️ Issues Discovered
-- [timestamp] move_note: Slow with deep folder paths (2.1s) #performance
-- [timestamp] search_notes: Unicode query returned unexpected results #bug
-- [timestamp] project switch: Context lost for memory:// URLs #issue
+- [timestamp] ⚠️ move_note: Slow with deep folder paths (2.1s) #tier2 #performance
+- [timestamp] 🚨 search_notes: Unicode query returned unexpected results #tier1 #bug #critical
+- [timestamp] ⚠️ build_context: Context lost for memory:// URLs #tier2 #issue
 
 ### 🚀 Enhancements Identified
 - edit_note could benefit from preview mode #ux-improvement
@@ -219,7 +409,7 @@ permalink: test-session-[phase]-[timestamp]
 ### 📊 Performance Metrics
 - Average write_note time: 0.3s
 - Search with 100+ notes: 0.6s
-- Project switch overhead: 0.1s
+- Project parameter overhead: <0.1s
 - Memory usage: [observed levels]
 
 ## Relations
@@ -239,7 +429,7 @@ permalink: test-session-[phase]-[timestamp]
 - Learning curve and intuitiveness
 
 **System Behavior:**
-- Context preservation across operations
+- Stateless operation independence
 - memory:// URL navigation reliability
 - Multi-step workflow cohesion
 - Edge case graceful handling
@@ -353,24 +543,44 @@ For each error discovered:
    - Create comprehensive summary report
    - Generate development recommendations
 
+## Testing Success Criteria
+
+### Core Testing (Tier 1) - Must Pass
+- All 6 critical tools function correctly
+- No critical bugs in essential workflows
+- Acceptable performance for basic operations
+- Error handling works as expected
+
+### Feature Testing (Tier 1+2) - Should Pass
+- All 11 core + important tools function
+- Workflow scenarios complete successfully
+- Performance meets baseline expectations
+- Integration points work correctly
+
+### Comprehensive Testing (All Tiers) - Complete Coverage
+- All tools tested across all scenarios
+- Edge cases and stress testing completed
+- Performance baselines established
+- Full documentation of issues and enhancements
+
 ## Expected Outcomes
 
 **System Validation:**
-- v0.13.0 feature verification in real usage
-- Edge case discovery beyond unit tests
+- Feature verification prioritized by tier importance
+- Recent changes validated for regression
 - Performance baseline establishment
-- Bug identification with reproduction cases
+- Bug identification with severity assessment
 
 **Knowledge Base Creation:**
-- Comprehensive testing documentation
+- Prioritized testing documentation
 - Real usage examples for user guides
-- Edge case scenarios for future testing
+- Recent changes validation records
 - Performance insights for optimization
 
 **Development Insights:**
-- Prioritized bug fix list
+- Tier-based bug priority list
+- Recent changes impact assessment
 - Enhancement ideas from real usage
-- Architecture validation results
 - User experience improvement areas
 
 ## Post-Test Deliverables
@@ -402,9 +612,11 @@ For each error discovered:
    - Add performance benchmarks and targets
 
 ## Context
-- Uses installed basic-memory version (not development)
+- Uses real installed basic-memory version 
 - Tests complete MCP→API→DB→File stack
 - Creates living documentation in Basic Memory itself
 - Follows integration over isolation philosophy
+- Prioritizes testing by tool importance and usage frequency
+- Adapts to recent development changes dynamically
 - Focuses on real usage patterns over checklist validation
-- Generates actionable insights for development team
+- Generates actionable insights prioritized by impact

.claude/settings.json

@@ -0,0 +1,5 @@
+{
+  "enabledPlugins": {
+    "basic-memory@basicmachines": true
+  }
+}

.dockerignore

@@ -0,0 +1,60 @@
+# Git files
+.git/
+.gitignore
+.gitattributes
+
+# Development files
+.vscode/
+.idea/
+*.swp
+*.swo
+*~
+
+# Testing files
+tests/
+test-int/
+.pytest_cache/
+.coverage
+htmlcov/
+
+# Build artifacts
+build/
+dist/
+*.egg-info/
+__pycache__/
+*.pyc
+*.pyo
+*.pyd
+.Python
+
+# Virtual environments (uv creates these during build)
+.venv/
+venv/
+.env
+
+# CI/CD files
+.github/
+
+# Documentation (keep README.md and pyproject.toml)
+docs/
+CHANGELOG.md
+CLAUDE.md
+CONTRIBUTING.md
+
+# Example files not needed for runtime
+examples/
+
+# Local development files
+.basic-memory/
+*.db
+*.sqlite3
+
+# OS files
+.DS_Store
+Thumbs.db
+
+# Temporary files
+tmp/
+temp/
+*.tmp
+*.log

.env.example

@@ -0,0 +1,28 @@
+# 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

.env.oauth.example

@@ -1,55 +0,0 @@
-# OAuth Configuration for Basic Memory MCP Server
-# Copy this file to .env and update the values
-
-# Enable OAuth authentication
-FASTMCP_AUTH_ENABLED=true
-
-# OAuth provider type: basic, github, google, or supabase
-# - basic: Built-in OAuth provider with in-memory storage
-# - github: Integrate with GitHub OAuth
-# - google: Integrate with Google OAuth
-# - supabase: Integrate with Supabase Auth (recommended for production)
-FASTMCP_AUTH_PROVIDER=basic
-
-# OAuth issuer URL (your MCP server URL)
-FASTMCP_AUTH_ISSUER_URL=http://localhost:8000
-
-# Documentation URL for OAuth endpoints
-FASTMCP_AUTH_DOCS_URL=http://localhost:8000/docs/oauth
-
-# Required scopes (comma-separated)
-# Examples: read,write,admin
-FASTMCP_AUTH_REQUIRED_SCOPES=read,write
-
-# Secret key for JWT tokens (auto-generated if not set)
-# FASTMCP_AUTH_SECRET_KEY=your-secret-key-here
-
-# Enable client registration endpoint
-FASTMCP_AUTH_CLIENT_REGISTRATION_ENABLED=true
-
-# Enable token revocation endpoint  
-FASTMCP_AUTH_REVOCATION_ENABLED=true
-
-# Default scopes for new clients
-FASTMCP_AUTH_DEFAULT_SCOPES=read
-
-# Valid scopes that can be requested
-FASTMCP_AUTH_VALID_SCOPES=read,write,admin
-
-# Client secret expiry in seconds (optional)
-# FASTMCP_AUTH_CLIENT_SECRET_EXPIRY=86400
-
-# GitHub OAuth settings (if using github provider)
-# GITHUB_CLIENT_ID=your-github-client-id
-# GITHUB_CLIENT_SECRET=your-github-client-secret
-
-# Google OAuth settings (if using google provider)
-# GOOGLE_CLIENT_ID=your-google-client-id  
-# GOOGLE_CLIENT_SECRET=your-google-client-secret
-
-# Supabase settings (if using supabase provider)
-# SUPABASE_URL=https://your-project.supabase.co
-# SUPABASE_ANON_KEY=your-anon-key
-# SUPABASE_SERVICE_KEY=your-service-key  # Optional, for admin operations
-# SUPABASE_JWT_SECRET=your-jwt-secret   # Optional, for token validation
-# SUPABASE_ALLOWED_CLIENTS=client1,client2  # Comma-separated list of allowed client IDs

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

@@ -0,0 +1,83 @@
+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

@@ -0,0 +1,71 @@
+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

@@ -9,106 +9,60 @@ 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 == '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'
+      )
     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@beta
+        uses: anthropics/claude-code-action@v1
         with:
-          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
+          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:*)'
+

.github/workflows/docker.yml

@@ -0,0 +1,61 @@
+name: Docker Image CI
+
+on:
+  push:
+    tags:
+      - 'v*'  # Trigger on version tags like v1.0.0, v0.13.0, etc.
+  workflow_dispatch:  # Allow manual triggering for testing
+
+env:
+  REGISTRY: ghcr.io
+  IMAGE_NAME: basicmachines-co/basic-memory
+
+jobs:
+  docker:
+    runs-on: ubuntu-latest
+    permissions:
+      contents: read
+      packages: write
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+        with:
+          fetch-depth: 0
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+        with:
+          platforms: linux/amd64,linux/arm64
+
+      - name: Log in to GitHub Container Registry
+        uses: docker/login-action@v3
+        with:
+          registry: ${{ env.REGISTRY }}
+          username: ${{ github.actor }}
+          password: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Extract metadata
+        id: meta
+        uses: docker/metadata-action@v5
+        with:
+          images: ${{ env.REGISTRY }}/${{ env.IMAGE_NAME }}
+          tags: |
+            type=ref,event=branch
+            type=ref,event=pr
+            type=semver,pattern={{version}}
+            type=semver,pattern={{major}}.{{minor}}
+            type=raw,value=latest,enable={{is_default_branch}}
+
+      - name: Build and push Docker image
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: ./Dockerfile
+          platforms: linux/amd64,linux/arm64
+          push: true
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+          cache-from: type=gha
+          cache-to: type=gha,mode=max
+

.github/workflows/release.yml

@@ -51,4 +51,35 @@ jobs:
       - name: Publish to PyPI
         uses: pypa/gh-action-pypi-publish@release/v1
         with:
-          password: ${{ secrets.PYPI_TOKEN }}
+          password: ${{ secrets.PYPI_TOKEN }}
+
+  homebrew:
+    name: Update Homebrew Formula
+    needs: release
+    runs-on: ubuntu-latest
+    # Only run for stable releases (not dev, beta, or rc versions)
+    if: ${{ !contains(github.ref_name, 'dev') && !contains(github.ref_name, 'b') && !contains(github.ref_name, 'rc') }}
+    permissions:
+      contents: write
+      actions: read
+    steps:
+      - name: Update Homebrew formula
+        uses: mislav/bump-homebrew-formula-action@v3
+        with:
+          # Formula name in homebrew-basic-memory repo
+          formula-name: basic-memory
+          # The tap repository
+          homebrew-tap: basicmachines-co/homebrew-basic-memory
+          # Base branch of the tap repository
+          base-branch: main
+          # Download URL will be automatically constructed from the tag
+          download-url: https://github.com/basicmachines-co/basic-memory/archive/refs/tags/${{ github.ref_name }}.tar.gz
+          # Commit message for the formula update
+          commit-message: |
+            {{formulaName}} {{version}}
+
+            Created by https://github.com/basicmachines-co/basic-memory/actions/runs/${{ github.run_id }}
+        env:
+          # Personal Access Token with repo scope for homebrew-basic-memory repo
+          COMMITTER_TOKEN: ${{ secrets.HOMEBREW_TOKEN }}
+

.github/workflows/test.yml

@@ -13,12 +13,73 @@ on:
     branches: [ "main" ]
 
 jobs:
-  test:
-    runs-on: ubuntu-latest
+  test-sqlite:
+    name: Test SQLite (${{ matrix.os }}, Python ${{ matrix.python-version }})
     strategy:
       fail-fast: false
       matrix:
-        python-version: [ "3.12" ]
+        os: [ubuntu-latest, windows-latest]
+        python-version: [ "3.12", "3.13", "3.14" ]
+    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
+
+      - name: Install just (Linux)
+        if: runner.os != 'Windows'
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y just
+
+      - name: Install just (Windows)
+        if: runner.os == 'Windows'
+        run: |
+          # Install just using Chocolatey (pre-installed on GitHub Actions Windows runners)
+          choco install just --yes
+        shell: pwsh
+
+      - name: Create virtual env
+        run: |
+          uv venv
+
+      - name: Install dependencies
+        run: |
+          uv pip install -e ".[dev,semantic]"
+
+      - name: Run type checks
+        run: |
+          just typecheck
+
+      - name: Run linting
+        run: |
+          just lint
+
+      - name: Run tests (SQLite)
+        run: |
+          uv pip install pytest pytest-cov
+          just test-sqlite
+
+  test-postgres:
+    name: Test Postgres (Python ${{ matrix.python-version }})
+    strategy:
+      fail-fast: false
+      matrix:
+        python-version: [ "3.12", "3.13", "3.14" ]
+    runs-on: ubuntu-latest
+
+    # Note: No services section needed - testcontainers handles Postgres in Docker
 
     steps:
       - uses: actions/checkout@v4
@@ -37,7 +98,8 @@ jobs:
 
       - name: Install just
         run: |
-          curl --proto '=https' --tlsv1.2 -sSf https://just.systems/install.sh | bash -s -- --to /usr/local/bin
+          sudo apt-get update
+          sudo apt-get install -y just
 
       - name: Create virtual env
         run: |
@@ -45,13 +107,64 @@ jobs:
 
       - name: Install dependencies
         run: |
-          uv pip install -e .[dev]
+          uv pip install -e ".[dev,semantic]"
 
-      - name: Run type checks
-        run: |
-          just type-check
-
-      - name: Run tests
+      - name: Run tests (Postgres via testcontainers)
         run: |
           uv pip install pytest pytest-cov
-          just test
+          just test-postgres
+
+  coverage:
+    name: Coverage Summary (combined, Python 3.12)
+    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
+
+      - name: Install just
+        run: |
+          sudo apt-get update
+          sudo apt-get install -y just
+
+      - name: Create virtual env
+        run: |
+          uv venv
+
+      - name: Install dependencies
+        run: |
+          uv pip install -e ".[dev,semantic]"
+
+      - name: Run combined coverage (SQLite + Postgres)
+        run: |
+          uv pip install pytest pytest-cov
+          just coverage
+
+      - name: Add coverage report to job summary
+        if: always()
+        run: |
+          {
+            echo "## Coverage"
+            echo ""
+            echo '```'
+            uv run coverage report -m
+            echo '```'
+          } >> "$GITHUB_STEP_SUMMARY"
+
+      - name: Upload HTML coverage report
+        if: always()
+        uses: actions/upload-artifact@v4
+        with:
+          name: htmlcov
+          path: htmlcov/

.gitignore

@@ -1,6 +1,7 @@
 *.py[cod]
 __pycache__/
 .pytest_cache/
+.testmondata*
 .coverage
 htmlcov/
 
@@ -52,4 +53,8 @@ ENV/
 
 # claude action
 claude-output
-**/.claude/settings.local.json
+**/.claude/settings.local.json
+.mcp.json
+.mcpregistry_*
+/.testmondata
+.benchmarks/

.mcp.json

@@ -1,14 +0,0 @@
-{
-  "mcpServers": {
-    "basic-memory": {
-      "command": "uv",
-      "args": [
-        "--directory",
-        "/Users/phernandez/dev/basicmachines/basic-memory",
-        "run",
-        "src/basic_memory/cli/main.py",
-        "mcp"
-      ]
-    }
-  }
-}

.python-version

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

AGENTS.md

@@ -0,0 +1,445 @@
+# 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`
+- 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.

AUTH.md

@@ -1,42 +0,0 @@
-# OAuth Quick Start
-
-Basic Memory supports OAuth authentication for secure access control. For detailed documentation, see [OAuth Authentication Guide](docs/OAuth%20Authentication%20Guide.md).
-
-## Quick Test with MCP Inspector
-
-```bash
-# 1. Set a consistent secret key
-export FASTMCP_AUTH_SECRET_KEY="test-secret-key"
-
-# 2. Start server with OAuth
-FASTMCP_AUTH_ENABLED=true basic-memory mcp --transport streamable-http
-
-# 3. In another terminal, get a test token
-export FASTMCP_AUTH_SECRET_KEY="test-secret-key"  # Same key!
-basic-memory auth test-auth
-
-# 4. Copy the access token and use in MCP Inspector:
-#    - Server URL: http://localhost:8000/mcp 
-#    - Transport: streamable-http
-#    - Custom Headers:
-#      Authorization: Bearer YOUR_ACCESS_TOKEN
-#      Accept: application/json, text/event-stream
-```
-
-## OAuth Endpoints
-
-- `GET /authorize` - Authorization endpoint
-- `POST /token` - Token exchange endpoint
-- `GET /.well-known/oauth-authorization-server` - OAuth metadata
-
-## Common Issues
-
-1. **401 Unauthorized**: Make sure you're using the same secret key for both server and client
-2. **404 Not Found**: Use `/authorize` not `/auth/authorize`
-3. **Token Invalid**: Tokens don't persist across server restarts with basic provider
-
-## Documentation
-
-- [OAuth Authentication Guide](docs/OAuth%20Authentication%20Guide.md) - Complete setup guide
-- [Supabase OAuth Setup](docs/Supabase%20OAuth%20Setup.md) - Production deployment
-- [External OAuth Providers](docs/External%20OAuth%20Providers.md) - GitHub, Google integration

CLA.md

@@ -1,34 +1,71 @@
-Developer Certificate of Origin
-Version 1.1
-https://developercertificate.org/
+# Contributor License Agreement
 
-Copyright (C) 2004, 2006 The Linux Foundation and its contributors.
+## Copyright Assignment and License Grant
 
-Everyone is permitted to copy and distribute verbatim copies of this
-license document, but changing it is not allowed.
+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.
 
-Developer's Certificate of Origin 1.1
+### 1. Definitions
 
-By making a contribution to this project, I certify that:
+"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.
 
-(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
+"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").
 
-(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
+### 2. Grant of Copyright License
 
-(c) The contribution was provided directly to me by some other
-person who certified (a), (b) or (c) and I have not modified
-it.
+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.
 
-(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.
+### 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.

CLAUDE.md

@@ -1,224 +0,0 @@
-# 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
-    - `read_file(path)` - Read raw file content (text, images, binaries) without knowledge graph processing
-
-  **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")
-
-  **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 (Manual)
-- Create version tag: `git tag v0.13.0 && git push origin v0.13.0`
-- Automatically builds, creates GitHub release, and publishes to PyPI
-- Users install with: `pip install basic-memory`
-
-### For Development
-- No manual version bumping required
-- Versions automatically derived from git tags
-- `pyproject.toml` uses `dynamic = ["version"]`
-- `__init__.py` dynamically reads version from package metadata

CLAUDE.md

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

CONTRIBUTING.md

@@ -27,13 +27,25 @@ 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. **Run the Tests**:
+3. **Activate the Virtual Environment**
    ```bash
-   # Run all tests
+   source .venv/bin/activate
+   ```
+
+4. **Run the Tests**:
+   ```bash
+   # Run all tests with unified coverage (unit + integration)
    just test
-   # or
-   uv run pytest -p pytest_mock -v
-   
+
+   # 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
+
    # Run a specific test
    pytest tests/path/to/test_file.py::test_function_name
    ```
@@ -129,7 +141,7 @@ agreement to the DCO.
 
 ## Code Style Guidelines
 
-- **Python Version**: Python 3.12+ with full type annotations
+- **Python Version**: Python 3.12+ with full type annotations (3.12+ required for type parameter syntax)
 - **Line Length**: 100 characters maximum
 - **Formatting**: Use ruff for consistent styling
 - **Import Order**: Standard lib, third-party, local imports
@@ -139,12 +151,78 @@ agreement to the DCO.
 
 ## Testing Guidelines
 
-- **Coverage Target**: We aim for 100% test coverage for all code
+### 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
 - **Test Framework**: Use pytest for unit and integration tests
-- **Mocking**: Use pytest-mock for mocking dependencies only when necessary
+- **Mocking**: Avoid mocking in integration tests; use sparingly in unit tests
 - **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

@@ -1,16 +1,52 @@
-# Generated by https://smithery.ai. See: https://smithery.ai/docs/config#dockerfile
-FROM python:3.12-slim
+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
+
+# Copy the project into the image
+ADD . /app
+
+# Install Python 3.13 explicitly and sync the project
 WORKDIR /app
+RUN uv python install 3.13
+RUN uv sync --locked --python 3.13
 
-# Copy the project files
-COPY . .
+# Create necessary directories and set ownership
+RUN mkdir -p /app/data/basic-memory /app/.basic-memory && \
+    chown -R appuser:${GID} /app
 
-# Install pip and build dependencies
-RUN pip install --upgrade pip \
-&& pip install . --no-cache-dir --ignore-installed
+# Set default data directory and add venv to PATH
+ENV BASIC_MEMORY_HOME=/app/data/basic-memory \
+    BASIC_MEMORY_PROJECT_ROOT=/app/data \
+    PATH="/app/.venv/bin:$PATH"
 
-# Expose port if necessary (e.g., uv might use a port, but MCP over stdio so not needed here)
+# Switch to the non-root user
+USER appuser
 
-# Use the basic-memory entrypoint to run the MCP server
-CMD ["basic-memory", "mcp"]
+# Expose port
+EXPOSE 8000
+
+# Health check
+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"]

NOTE-FORMAT.md

@@ -0,0 +1,494 @@
+# 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

@@ -1,3 +1,4 @@
+<!-- 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/)
@@ -5,7 +6,16 @@
 [![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')
-[![smithery badge](https://smithery.ai/badge/@basicmachines-co/basic-memory)](https://smithery.ai/server/@basicmachines-co/basic-memory)
+
+## 🚀 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 `{{OSS_DISCOUNT_CODE}}` for 20% off for 3 months.
+
+[Sign up now →](https://basicmemory.com)
+
+with a 7 day free trial
 
 # Basic Memory
 
@@ -13,8 +23,8 @@ 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.
 
-- Website: https://basicmachines.co
-- Documentation: https://memory.basicmachines.co
+- Website: https://basicmemory.com
+- Documentation: https://docs.basicmemory.com
 
 ## Pick up your conversation right where you left off
 
@@ -52,24 +62,6 @@ uv tool install basic-memory
 
 You can view shared context via files in `~/basic-memory` (default directory location).
 
-### Alternative Installation via Smithery
-
-You can use [Smithery](https://smithery.ai/server/@basicmachines-co/basic-memory) to automatically configure Basic
-Memory for Claude Desktop:
-
-```bash
-npx -y @smithery/cli install @basicmachines-co/basic-memory --client claude
-```
-
-This installs and configures Basic Memory without requiring manual edits to the Claude Desktop configuration file. The
-Smithery server hosts the MCP server component, while your data remains stored locally as Markdown files.
-
-### Glama.ai
-
-<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?
 
 Most LLM interactions are ephemeral - you ask a question, get an answer, and everything is forgotten. Each conversation
@@ -100,6 +92,9 @@ 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
 
@@ -153,7 +148,7 @@ 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 is enabled by default with the v0.12.0 version
+- Realtime sync can be enabled via running `basic-memory sync --watch`
 
 4. In a chat with the LLM, you can reference a topic:
 
@@ -263,13 +258,6 @@ 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)`.
 
@@ -299,6 +287,8 @@ 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/):
@@ -322,8 +312,7 @@ for OS X):
 }
 ```
 
-If you want to use a specific project (see [Multiple Projects](docs/User%20Guide.md#multiple-projects)), update your
-Claude Desktop
+If you want to use a specific project (see [Multiple Projects](#multiple-projects) below), update your Claude Desktop
 config:
 
 ```json
@@ -333,9 +322,9 @@ config:
       "command": "uvx",
       "args": [
         "basic-memory",
+        "mcp",
         "--project",
-        "your-project-name",
-        "mcp"
+        "your-project-name"
       ]
     }
   }
@@ -344,16 +333,109 @@ config:
 
 2. Sync your knowledge:
 
-Basic Memory will sync the files in your project in real time if you make manual edits.
+```bash
+# One-time sync of local knowledge updates
+basic-memory sync
 
-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 (global cloud mode via OAuth)
+basic-memory cloud login
+
+# Bidirectional sync with cloud
+basic-memory cloud sync
+
+# Verify cloud integrity
+basic-memory cloud check
+
+# Mount cloud storage
+basic-memory cloud mount
+```
+
+**Per-Project Cloud Routing** (API key based):
+
+Individual projects can be routed through the cloud while others stay local. This uses an API key instead of OAuth:
+
+```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 with mode column (local/cloud)
+basic-memory project list
+```
+
+**Routing Flags** (for users with global cloud mode):
+
+When global cloud mode is enabled, CLI commands communicate with the cloud API by default. Use routing flags to override this:
+
+```bash
+# Force local routing (useful for local MCP server while cloud mode is enabled)
+basic-memory status --local
+basic-memory project list --local
+
+# Force cloud routing (when cloud mode is disabled but you want cloud access)
+basic-memory status --cloud
+basic-memory project info my-project --cloud
+```
+
+The local MCP server (`basic-memory mcp`) automatically uses local routing, so you can use both local Claude Desktop and cloud-based clients simultaneously.
+
+4. In Claude Desktop, the LLM can now use these tools:
+
+**Content Management:**
 ```
 write_note(title, content, folder, tags) - Create or update notes
 read_note(identifier, page, page_size) - 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) - Edit notes incrementally
+move_note(identifier, destination_path) - Move notes with database consistency
+delete_note(identifier) - Delete notes from knowledge base
+```
+
+**Knowledge Graph Navigation:**
+```
 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
+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
+search_by_metadata(filters, limit, offset, project) - Structured frontmatter search
+```
+
+**Project Management:**
+```
+list_memory_projects() - List all available projects
+create_memory_project(project_name, project_path) - Create new projects
+get_current_project() - Show current project stats
+sync_status() - Check synchronization status
+```
+
+**Cloud Discovery (opt-in):**
+```
+cloud_info() - Show optional Cloud overview and setup guidance
+release_notes() - Show latest release notes
+```
+
+**Visualization:**
+```
 canvas(nodes, edges, title, folder) - Generate knowledge visualizations
 ```
 
@@ -369,31 +451,114 @@ canvas(nodes, edges, title, folder) - Generate knowledge visualizations
 
 ## Futher info
 
-See the [Documentation](https://memory.basicmachines.co/) for more info, including:
+See the [Documentation](https://docs.basicmemory.com) for more info, including:
 
-- [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)
+- [Complete User Guide](https://docs.basicmemory.com/user-guide/)
+- [CLI tools](https://docs.basicmemory.com/guides/cli-reference/)
+- [Cloud CLI and Sync](https://docs.basicmemory.com/guides/cloud-cli/)
+- [Managing multiple Projects](https://docs.basicmemory.com/guides/cli-reference/#project)
+- [Importing data from OpenAI/Claude Projects](https://docs.basicmemory.com/guides/cli-reference/#import)
 
-## Installation Options
+## 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 (ignores cloud mode) |
+| `BASIC_MEMORY_ENV` | `dev` | Set to `test` for test mode (stderr only) |
+
+### Examples
 
-### Stable Release
 ```bash
-pip install basic-memory
+# 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
 ```
 
-### Beta/Pre-releases
+## 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:**
 ```bash
-pip install basic-memory --pre
+# 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
 ```
 
-### Development Builds
-Development versions are automatically published on every commit to main with versions like `0.12.4.dev26+468a22f`:
+**Available Test Commands:**
+
+- `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
+
+**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
-pip install basic-memory --pre --force-reinstall
+just install          # Install with dev dependencies
+just lint             # Run linting checks
+just typecheck        # Run type checking
+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
 ```
 
+**Local Consistency Check:**
+```bash
+basic-memory doctor   # Verifies file <-> database sync in a temp project
+```
+
+See the [justfile](justfile) for the complete list of development commands.
+
 ## License
 
 AGPL-3.0

RELEASE_NOTES_v0.13.0.md

@@ -1,241 +0,0 @@
-# Release Notes v0.13.0
-
-## Overview
-
-Basic Memory v0.13.0 is a **major release** that transforms Basic Memory into a true multi-project knowledge management system. This release introduces fluid project switching, advanced note editing capabilities, robust file management, and production-ready OAuth authentication - all while maintaining full backward compatibility.
-
-**What's New for Users:**
-- 🎯 **Switch between projects instantly** during conversations with Claude
-- ✏️ **Edit notes incrementally** without rewriting entire documents
-- 📁 **Move and organize notes** with full database consistency
-- 📖 **View notes as formatted artifacts** for better readability in Claude Desktop
-- 🔍 **Search frontmatter tags** to discover content more easily
-- 🔐 **OAuth authentication** for secure remote access
-- ⚡ **Development builds** automatically published for beta testing
-
-**Key v0.13.0 Accomplishments:**
-- ✅ **Complete Project Management System** - Project switching and project-specific operations
-- ✅ **Advanced Note Editing** - Incremental editing with append, prepend, find/replace, and section operations  
-- ✅ **View Notes as Artifacts in Claude Desktop/Web** - Use the view_note tool to view a note as an artifact
-- ✅ **File Management System** - Full move operations with database consistency and rollback protection
-- ✅ **Enhanced Search Capabilities** - Frontmatter tags now searchable, improved content discoverability
-- ✅ **Unified Database Architecture** - Single app-level database for better performance and project management
-
-## Major Features
-
-### 1. Multiple Project Management 🎯
-
-**Switch between projects instantly during conversations:**
-
-```
-💬 "What projects do I have?"
-🤖 Available projects:
-   • main (current, default)
-   • work-notes
-   • personal-journal
-   • code-snippets
-
-💬 "Switch to work-notes"
-🤖 ✓ Switched to work-notes project
-   
-   Project Summary:
-   • 47 entities
-   • 125 observations  
-   • 23 relations
-
-💬 "What did I work on yesterday?"
-🤖 [Shows recent activity from work-notes project]
-```
-
-**Key Capabilities:**
-- **Instant Project Switching**: Change project context mid-conversation without restart
-- **Project-Specific Operations**: Operations work within the currently active project context
-- **Project Discovery**: List all available projects with status indicators
-- **Session Context**: Maintains active project throughout conversation
-- **Backward Compatibility**: Existing single-project setups continue to work seamlessly
-
-### 2. Advanced Note Editing ✏️
-
-**Edit notes incrementally without rewriting entire documents:**
-
-```python
-# Append new sections to existing notes
-edit_note("project-planning", "append", "\n## New Requirements\n- Feature X\n- Feature Y")
-
-# Prepend timestamps to meeting notes
-edit_note("meeting-notes", "prepend", "## 2025-05-27 Update\n- Progress update...")
-
-# Replace specific sections under headers
-edit_note("api-spec", "replace_section", "New implementation details", section="## Implementation")
-
-# Find and replace with validation
-edit_note("config", "find_replace", "v0.13.0", find_text="v0.12.0", expected_replacements=2)
-```
-
-**Key Capabilities:**
-- **Append Operations**: Add content to end of notes (most common use case)
-- **Prepend Operations**: Add content to beginning of notes
-- **Section Replacement**: Replace content under specific markdown headers
-- **Find & Replace**: Simple text replacements with occurrence counting
-- **Smart Error Handling**: Helpful guidance when operations fail
-- **Project Context**: Works within the active project with session awareness
-
-### 3. Smart File Management 📁
-
-**Move and organize notes:**
-
-```python
-# Simple moves with automatic folder creation
-move_note("my-note", "work/projects/my-note.md")
-
-# Organize within the active project
-move_note("shared-doc", "archive/old-docs/shared-doc.md")
-
-# Rename operations
-move_note("old-name", "same-folder/new-name.md")
-```
-
-**Key Capabilities:**
-- **Database Consistency**: Updates file paths, permalinks, and checksums automatically
-- **Search Reindexing**: Maintains search functionality after moves
-- **Folder Creation**: Automatically creates destination directories
-- **Project Isolation**: Operates within the currently active project
-- **Link Preservation**: Maintains internal links and references
-
-### 4. Enhanced Search & Discovery 🔍
-
-**Find content more easily with improved search capabilities:**
-
-- **Frontmatter Tag Search**: Tags from YAML frontmatter are now indexed and searchable
-- **Improved Content Discovery**: Search across titles, content, tags, and metadata
-- **Project-Scoped Search**: Search within the currently active project
-- **Better Search Quality**: Enhanced FTS5 indexing with tag content inclusion
-
-**Example:**
-```yaml
----
-title: Coffee Brewing Methods
-tags: [coffee, brewing, equipment]
----
-```
-Now searchable by: "coffee", "brewing", "equipment", or "Coffee Brewing Methods"
-
-### 5. Unified Database Architecture 🗄️
-
-**Single app-level database for better performance and project management:**
-
-- **Migration from Per-Project DBs**: Moved from multiple SQLite files to single app database
-- **Project Isolation**: Proper data separation with project_id foreign keys
-- **Better Performance**: Optimized queries and reduced file I/O
-
-## Complete MCP Tool Suite 🛠️
-
-### New Project Management Tools
-- **`list_projects()`** - Discover and list all available projects with status
-- **`switch_project(project_name)`** - Change active project context during conversations
-- **`get_current_project()`** - Show currently active project with statistics
-- **`set_default_project(project_name)`** - Update default project configuration
-- **`sync_status()`** - Check file synchronization status and background operations
-
-### New Note Operations Tools
-- **`edit_note()`** - Incremental note editing (append, prepend, find/replace, section replace)
-- **`move_note()`** - Move notes with database consistency and search reindexing
-- **`view_note()`** - Display notes as formatted artifacts for better readability in Claude Desktop
-
-### Enhanced Existing Tools
-All existing tools now support:
-- **Session context awareness** (operates within the currently active project)
-- **Enhanced error messages** with project context metadata
-- **Improved response formatting** with project information footers
-- **Project isolation** ensures operations stay within the correct project boundaries
-
-
-## User Experience Improvements
-
-### Installation Options
-
-**Multiple ways to install and test Basic Memory:**
-
-```bash
-# Stable release
-uv tool install basic-memory
-
-# Beta/pre-releases
-uv tool install basic-memory --pre
-```
-
-
-### Bug Fixes & Quality Improvements
-
-**Major issues resolved in v0.13.0:**
-
-- **#118**: Fixed YAML tag formatting to follow standard specification
-- **#110**: Fixed `--project` flag consistency across all CLI commands
-- **#107**: Fixed write_note update failures with existing notes
-- **#93**: Fixed custom permalink handling in frontmatter
-- **#52**: Enhanced search capabilities with frontmatter tag indexing
-- **FTS5 Search**: Fixed special character handling in search queries
-- **Error Handling**: Improved error messages and validation across all tools
-
-## Breaking Changes & Migration
-
-### For Existing Users
-
-**Automatic Migration**: First run will automatically migrate existing data to the new unified database structure. No manual action required.
-
-**What Changes:**
-- Database location: Moved to `~/.basic-memory/memory.db` (unified across projects)
-- Configuration: Projects defined in `~/.basic-memory/config.json` are synced with database
-
-**What Stays the Same:**
-- All existing notes and data remain unchanged
-- Default project behavior maintained for single-project users
-- All existing MCP tools continue to work without modification
-
-
-
-
-## Documentation & Resources
-
-### New Documentation
-- [Project Management Guide](docs/Project%20Management.md) - Multi-project workflows
-- [Note Editing Guide](docs/Note%20Editing.md) - Advanced editing techniques
-
-### Updated Documentation
-- [README.md](README.md) - Installation options and beta build instructions
-- [CONTRIBUTING.md](CONTRIBUTING.md) - Release process and version management
-- [CLAUDE.md](CLAUDE.md) - Development workflow and CI/CD documentation
-- [Claude.ai Integration](docs/Claude.ai%20Integration.md) - Updated MCP tool examples
-
-### Quick Start Examples
-
-**Project Switching:**
-```
-💬 "Switch to my work project and show recent activity"
-🤖 [Calls switch_project("work") then recent_activity()]
-```
-
-**Note Editing:**
-```
-💬 "Add a section about deployment to my API docs"
-🤖 [Calls edit_note("api-docs", "append", "## Deployment\n...")]
-```
-
-**File Organization:**
-```
-💬 "Move my old meeting notes to the archive folder"
-🤖 [Calls move_note("meeting-notes", "archive/old-meetings.md")]
-```
-
-
-### Getting Updates
-```bash
-# Stable releases
-uv tool upgrade basic-memory
-
-# Beta releases  
-uv tool install basic-memory --pre --force-reinstall
-
-# Latest development
-uv tool install basic-memory --pre --force-reinstall
-```

docker-compose-postgres.yml

@@ -0,0 +1,42 @@
+# 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

docker-compose.yml

@@ -0,0 +1,83 @@
+# Docker Compose configuration for Basic Memory
+# See docs/Docker.md for detailed setup instructions
+
+version: '3.8'
+
+services:
+  basic-memory:
+    # Use pre-built image (recommended for most users)
+    image: ghcr.io/basicmachines-co/basic-memory:latest
+    
+    # Uncomment to build locally instead:
+    # build: .
+    
+    container_name: basic-memory-server
+    
+    # Volume mounts for knowledge directories and persistent data
+    volumes:
+
+      # Persistent storage for configuration and database
+      - basic-memory-config:/root/.basic-memory:rw
+
+      # Mount your knowledge directory (required)
+      # Change './knowledge' to your actual Obsidian vault or knowledge directory
+      - ./knowledge:/app/data:rw
+
+      # OPTIONAL: Mount additional knowledge directories for multiple projects
+      # - ./work-notes:/app/data/work:rw
+      # - ./personal-notes:/app/data/personal:rw
+
+      # You can edit the project config manually in the mounted config volume
+      # The default project will be configured to use /app/data
+    environment:
+      # Project configuration
+      - BASIC_MEMORY_DEFAULT_PROJECT=main
+      
+      # Enable real-time file synchronization (recommended for Docker)
+      - BASIC_MEMORY_SYNC_CHANGES=true
+      
+      # Logging configuration
+      - BASIC_MEMORY_LOG_LEVEL=INFO
+      
+      # Sync delay in milliseconds (adjust for performance vs responsiveness)
+      - BASIC_MEMORY_SYNC_DELAY=1000
+    
+    # Port exposure for HTTP transport (only needed if not using STDIO)
+    ports:
+      - "8000:8000"
+    
+    # Command with SSE transport (configurable via environment variables above)
+    # IMPORTANT: The SSE and streamable-http endpoints are not secured
+    command: ["basic-memory", "mcp", "--transport", "sse", "--host", "0.0.0.0", "--port", "8000"]
+    
+    # Container management
+    restart: unless-stopped
+    
+    # Health monitoring
+    healthcheck:
+      test: ["CMD", "basic-memory", "--version"]
+      interval: 30s
+      timeout: 10s
+      retries: 3
+      start_period: 30s
+    
+    # Optional: Resource limits
+    # deploy:
+    #   resources:
+    #     limits:
+    #       memory: 512M
+    #       cpus: '0.5'
+    #     reservations:
+    #       memory: 256M
+    #       cpus: '0.25'
+
+volumes:
+  # Named volume for persistent configuration and database
+  # This ensures your configuration and knowledge graph persist across container restarts
+  basic-memory-config:
+    driver: local
+
+# Network configuration (optional)
+# networks:
+#   basic-memory-net:
+#     driver: bridge

docs/AI Assistant Guide.md

@@ -1,431 +0,0 @@
----
-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

@@ -0,0 +1,444 @@
+# 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 (cloud/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(
+            cloud_mode_enabled=config.cloud_mode_enabled,
+            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: **TEST > CLOUD > LOCAL**
+
+```python
+def resolve_runtime_mode(cloud_mode_enabled: bool, is_test_env: bool) -> RuntimeMode:
+    if is_test_env:
+        return RuntimeMode.TEST
+    if cloud_mode_enabled:
+        return RuntimeMode.CLOUD
+    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` in config, which affects client routing in `get_client(project_name=...)` without changing the global runtime mode.
+
+## 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/CLI Reference.md

@@ -1,414 +0,0 @@
----
-title: CLI Reference
-type: note
-permalink: docs/cli-reference
----
-
-# CLI Reference
-
-Basic Memory provides command line tools for managing your knowledge base. This reference covers the available commands and their options.
-
-## Core Commands
-
-### auth (New in v0.13.0)
-
-Manage OAuth authentication for secure remote access:
-
-```bash
-# Test authentication setup
-basic-memory auth test-auth
-
-# Register OAuth client
-basic-memory auth register-client
-```
-
-Supports multiple authentication providers:
-- **Basic Provider**: For development and testing
-- **Supabase Provider**: For production deployments
-- **External Providers**: GitHub, Google integration framework
-
-See [[OAuth Authentication Guide]] for complete setup instructions.
-
-### sync
-
-Keeps files and the knowledge graph in sync:
-
-```bash
-# Basic sync
-basic-memory sync
-
-# Watch for changes
-basic-memory sync --watch
-
-# Show detailed sync information
-basic-memory sync --verbose
-```
-
-Options:
-- `--watch`: Continuously monitor for changes
-- `--verbose`: Show detailed output
-
-**Note**:
-
-As of the v0.12.0 release syncing will occur in real time when the mcp process starts.
-- The real time sync means that it is no longer necessary to run the `basic-memory sync --watch` process in a a terminal to sync changes to the db (so the AI can see them). This will be done automatically.
-
-This behavior can be changed via the config. The config file for Basic Memory is in the home directory under `.basic-memory/config.json`.
-
-To change the properties, set the following values:
-```
- ~/.basic-memory/config.json 
-{
-  "sync_changes": false,
-}
-```
-
-Thanks for using Basic Memory!
-### import (Enhanced in v0.13.0)
-
-Imports external knowledge sources with support for project targeting:
-
-```bash
-# Claude conversations
-basic-memory import claude conversations
-
-# Claude projects
-basic-memory import claude projects
-
-# ChatGPT history
-basic-memory import chatgpt
-
-# Memory JSON format
-basic-memory import memory-json /path/to/memory.json
-
-# Import to specific project (v0.13.0)
-basic-memory --project=work import claude conversations
-```
-
-**New in v0.13.0:**
-- **Project Targeting**: Import directly to specific projects
-- **Real-time Sync**: Imported content available immediately
-- **Unified Database**: All imports stored in centralized database
-
-> **Note**: Changes sync automatically - no manual sync required in v0.13.0.
-### status
-
-Shows system status information:
-
-```bash
-# Basic status check
-basic-memory status
-
-# Detailed status
-basic-memory status --verbose
-
-# JSON output
-basic-memory status --json
-```
-
-
-### project (Enhanced in v0.13.0)
-
-Manage multiple projects with the new unified database architecture. Projects can now be switched instantly during conversations without restart.
-  
-```bash  
-# List all configured projects with status
-basic-memory project list  
-  
-# Create a new project
-basic-memory project create work ~/work-basic-memory  
-  
-# Set the default project  
-basic-memory project set-default work  
-  
-# Delete a project (doesn't delete files)  
-basic-memory project delete personal  
-  
-# Show detailed project statistics
-basic-memory project info
-```  
-
-**New in v0.13.0:**
-- **Unified Database**: All projects share a single database for better performance
-- **Instant Switching**: Switch projects during conversations without restart
-- **Enhanced Commands**: Updated project commands with better status information
-- **Project Statistics**: Detailed info about entities, observations, and relations 
-
-#### Using Projects in Commands  
-  
-All commands support the `--project` flag to specify which project to use:  
-  
-```bash  
-# Sync a specific project  
-basic-memory --project=work sync  
-  
-# Run MCP server for a specific project  
-basic-memory --project=personal mcp  
-```  
-  
-You can also set the `BASIC_MEMORY_PROJECT` environment variable:  
-  
-```bash  
-BASIC_MEMORY_PROJECT=work basic-memory sync  
-```  
-
-### tool (Enhanced in v0.13.0)
-
-Direct access to MCP tools via CLI with new editing and file management capabilities:
-
-```bash
-# Create notes
-basic-memory tool write-note --title "My Note" --content "Content here"
-
-# Edit notes incrementally (v0.13.0)
-echo "New content" | basic-memory tool edit-note --title "My Note" --operation append
-
-# Move notes (v0.13.0)
-basic-memory tool move-note --identifier "My Note" --destination "archive/my-note.md"
-
-# Search notes
-basic-memory tool search-notes --query "authentication"
-
-# Project management (v0.13.0)
-basic-memory tool list-projects
-basic-memory tool switch-project --project-name "work"
-```
-
-**New in v0.13.0:**
-- **edit-note**: Incremental editing (append, prepend, find/replace, section replace)
-- **move-note**: File management with database consistency
-- **Project tools**: list-projects, switch-project, get-current-project
-- **Cross-project operations**: Use `--project` flag with any tool
-
-### help
-
-The full list of commands and help for each can be viewed with the `--help` argument.
-
-```
- ✗ basic-memory --help
-
- Usage: basic-memory [OPTIONS] COMMAND [ARGS]...
-
- Basic Memory - Local-first personal knowledge management system.
-
-╭─ Options ─────────────────────────────────────────────────────────────────────────────────╮
-│ --project             -p      TEXT  Specify which project to use                          │
-│                                     [env var: BASIC_MEMORY_PROJECT]                       │
-│                                     [default: None]                                       │
-│ --version             -V            Show version information and exit.                    │
-│ --install-completion                Install completion for the current shell.             │
-│ --show-completion                   Show completion for the current shell, to copy it or  │
-│                                     customize the installation.                           │
-│ --help                              Show this message and exit.                           │
-╰───────────────────────────────────────────────────────────────────────────────────────────╯
-╭─ Commands ────────────────────────────────────────────────────────────────────────────────╮
-│ auth      OAuth authentication management (v0.13.0)                                      │
-│ sync      Sync knowledge files with the database                                          │
-│ status    Show sync status between files and database                                     │
-│ reset     Reset database (drop all tables and recreate)                                   │
-│ mcp       Run the MCP server for Claude Desktop integration                               │
-│ import    Import data from various sources                                                │
-│ tool      Direct access to MCP tools via CLI                                              │
-│ project   Manage multiple Basic Memory projects                                           │
-╰───────────────────────────────────────────────────────────────────────────────────────────╯
-```
-
-## Initial Setup
-
-```bash
-# Install Basic Memory
-uv install basic-memory
-
-# First sync
-basic-memory sync
-
-# Start watching mode
-basic-memory sync --watch
-```
-
-> **Important**: You need to install Basic Memory via `uv` or `pip` to use the command line tools, see [[Getting Started with Basic Memory#Installation]].
-
-## Regular Usage
-
-```bash
-# Check status
-basic-memory status
-
-# Import new content
-basic-memory import claude conversations
-
-# Sync changes
-basic-memory sync
-
-# Sync changes continuously
-basic-memory sync --watch
-```
-
-## Maintenance Tasks
-
-```bash
-# Check system status in detail
-basic-memory status --verbose
-
-# Full resync of all files
-basic-memory sync
-
-# Import updates to specific folder
-basic-memory import claude conversations --folder new
-```
-
-
-## Using stdin with Basic Memory's `write_note` Tool
-
-The `write-note` tool supports reading content from standard input (stdin), allowing for more flexible workflows when creating or updating notes in your Basic Memory knowledge base.
-
-### Use Cases
-
-This feature is particularly useful for:
-
-1. **Piping output from other commands** directly into Basic Memory notes
-2. **Creating notes with multi-line content** without having to escape quotes or special characters
-3. **Integrating with AI assistants** like Claude Code that can generate content and pipe it to Basic Memory
-4. **Processing text data** from files or other sources
-
-### Basic Usage
-
-#### Method 1: Using a Pipe
-
-You can pipe content from another command into `write_note`:
-
-```bash
-# Pipe output of a command into a new note
-echo "# My Note\n\nThis is a test note" | basic-memory tool write-note --title "Test Note" --folder "notes"
-
-# Pipe output of a file into a new note
-cat README.md | basic-memory tool write-note --title "Project README" --folder "documentation"
-
-# Process text through other tools before saving as a note
-cat data.txt | grep "important" | basic-memory tool write-note --title "Important Data" --folder "data"
-```
-
-#### Method 2: Using Heredoc Syntax
-
-For multi-line content, you can use heredoc syntax:
-
-```bash
-# Create a note with heredoc
-cat << EOF | basic-memory tool write_note --title "Project Ideas" --folder "projects"
-# Project Ideas for Q2
-
-## AI Integration
-- Improve recommendation engine
-- Add semantic search to product catalog
-
-## Infrastructure
-- Migrate to Kubernetes
-- Implement CI/CD pipeline
-EOF
-```
-
-#### Method 3: Input Redirection
-
-You can redirect input from a file:
-
-```bash
-# Create a note from file content
-basic-memory tool write-note --title "Meeting Notes" --folder "meetings" < meeting_notes.md
-```
-
-## Integration with Claude Code
-
-This feature works well with Claude Code in the terminal:
-
-### cli
-
-In a Claude Code session, let Claude know he can use the basic-memory tools, then he can execute them via the cli:
-
-```
-⏺ Bash(echo "# Test Note from Claude\n\nThis is a test note created by Claude to test the stdin functionality." | basic-memory tool write-note --title "Claude Test Note" --folder "test" --tags "test" --tags "claude")…
-  ⎿  # Created test/Claude Test Note.md (23e00eec)
-     permalink: test/claude-test-note
-
-     ## Tags
-     - test, claude
-
-```
-
-### MCP
-
-Claude code can also now use mcp tools, so it can use any of the basic-memory tool natively. To install basic-memory in Claude Code:
-
-Run
-```
-claude mcp add basic-memory basic-memory mcp
-```
-
-For example: 
-
-```
-➜  ~ claude mcp add basic-memory basic-memory mcp
-Added stdio MCP server basic-memory with command: basic-memory mcp to project config
-➜  ~ claude mcp list
-basic-memory: basic-memory mcp
-```
-
-You can then use the `/mcp` command in the REPL:
-
-```
-/mcp
-  ⎿  MCP Server Status
-
-     • basic-memory: connected
-```
-
-## Version Management (New in v0.13.0)
-
-Basic Memory v0.13.0 introduces automatic version management and multiple installation options:
-
-```bash
-# Stable releases
-pip install basic-memory
-
-# Beta/pre-releases
-pip install basic-memory --pre
-
-# Latest development builds (auto-published)
-pip install basic-memory --pre --force-reinstall
-
-# Check current version
-basic-memory --version
-```
-
-**Version Types:**
-- **Stable**: `0.13.0` (manual git tags)
-- **Beta**: `0.13.0b1` (manual git tags) 
-- **Development**: `0.12.4.dev26+468a22f` (automatic from commits)
-
-## Troubleshooting Common Issues
-
-### Sync Conflicts
-
-If you encounter a file changed during sync error:
-1. Check the file referenced in the error message
-2. Resolve any conflicts manually
-3. Run sync again
-
-### Import Errors
-
-If import fails:
-1. Check that the source file is in the correct format
-2. Verify permissions on the target directory
-3. Use --verbose flag for detailed error information
-
-### Status Issues
-
-If status shows problems:
-1. Note any unresolved relations or warnings
-2. Run a full sync to attempt automatic resolution
-3. Check file permissions if database access errors occur
-
-
-## Relations
-- used_by [[Getting Started with Basic Memory]] (Installation instructions)
-- complements [[User Guide]] (How to use Basic Memory)
-- relates_to [[Introduction to Basic Memory]] (System overview)

docs/Canvas.md

@@ -1,107 +0,0 @@
----
-title: Canvas Visualizations
-type: note
-permalink: docs/canvas
-tags:
-- visualization
-- mapping
-- obsidian
----
-
-# Canvas Visualizations
-
-Basic Memory can create visual knowledge maps using Obsidian's Canvas feature. These visualizations help you understand relationships between concepts, map out processes, and visualize your knowledge structure.
-
-## Creating Canvas Visualizations
-
-Ask Claude to create a visualization by describing what you want to map:
-
-```
-You: "Create a canvas visualization of my project components and their relationships."
-
-You: "Make a concept map showing the main themes from our discussion about climate change."
-
-You: "Can you make a canvas diagram of the perfect pour over method?"
-```
-
-![[Canvas.png]]
-
-## Types of Visualizations
-
-Basic Memory can create several types of visual maps:
-
-### Document Maps
-Visualize connections between your notes and documents
-
-### Concept Maps
-Create visual representations of ideas and their relationships
-
-### Process Diagrams
-Map workflows, sequences, and procedures
-
-### Thematic Analysis
-Organize ideas around central themes
-
-### Relationship Networks
-Show how different entities relate to each other
-
-## Visualization Sources
-
-Claude can create visualizations based on:
-
-### Documents in Your Knowledge Base
-```
-You: "Create a canvas showing the connections between my project planning documents"
-```
-
-### Conversation Content
-```
-You: "Make a canvas visualization of the main points we just discussed"
-```
-
-### Search Results
-```
-You: "Find all my notes about psychology and create a visual map of the concepts"
-```
-
-### Themes and Relationships
-```
-You: "Create a visual map showing how different philosophical schools relate to each other"
-```
-
-## Visualization Workflow
-
-1. **Request a visualization** by describing what you want to see
-2. **Claude creates the canvas file** in your Basic Memory directory
-3. **Open the file in Obsidian** to view the visualization
-4. **Refine the visualization** by asking Claude for adjustments:
-   ```
-   You: "Could you reorganize the canvas to group related components together?"
-   
-   You: "Please add more detail about the connection between these two concepts."
-   ```
-
-## Technical Details
-
-Behind the scenes, Claude:
-
-1. Creates a `.canvas` file in JSON format
-2. Adds nodes for each concept or document
-3. Creates edges to represent relationships
-4. Sets positions for visual clarity
-5. Includes any relevant metadata
-
-The resulting file is fully compatible with Obsidian's Canvas feature and can be edited directly in Obsidian.
-
-## Tips for Effective Visualizations
-
-- **Be specific** about what you want to visualize
-- **Specify the level of detail** you need
-- **Mention the visualization type** you want (concept map, process flow, etc.)
-- **Start simple** and ask for refinements
-- **Provide context** about what documents or concepts to include
-
-## Relations
-- enhances [[Obsidian Integration]] (Using Basic Memory with Obsidian)
-- visualizes [[Knowledge Format]] (The structure of your knowledge)
-- complements [[User Guide]] (Ways to use Basic Memory)

docs/Claude.ai Integration.md

@@ -1,335 +0,0 @@
-# Claude.ai Integration Guide
-
-This guide explains how to connect Basic Memory to Claude.ai, enabling Claude to read and write to your personal knowledge base.
-
-## Overview
-
-When connected to Claude.ai, Basic Memory provides:
-- Persistent memory across conversations
-- Knowledge graph navigation
-- Note-taking and search capabilities
-- File organization and management
-
-## Prerequisites
-
-1. Basic Memory MCP server with OAuth enabled
-2. Public HTTPS URL (or tunneling service for testing)
-3. Claude.ai account (Free, Pro, or Enterprise)
-
-## Quick Start (Testing)
-
-### 1. Start MCP Server with OAuth
-
-```bash
-# Enable OAuth with basic provider
-export FASTMCP_AUTH_ENABLED=true
-export FASTMCP_AUTH_PROVIDER=basic
-
-# Start server on all interfaces
-basic-memory mcp --transport streamable-http --host 0.0.0.0 --port 8000
-```
-
-### 2. Make Server Accessible
-
-For testing, use ngrok:
-
-```bash
-# Install ngrok
-brew install ngrok  # macOS
-# or download from https://ngrok.com
-
-# Create tunnel
-ngrok http 8000
-```
-
-Note the HTTPS URL (e.g., `https://abc123.ngrok.io`)
-
-### 3. Register OAuth Client
-
-```bash
-# Register a client for Claude
-basic-memory auth register-client --client-id claude-ai
-
-# Save the credentials!
-# Client ID: claude-ai
-# Client Secret: xxx...
-```
-
-### 4. Connect in Claude.ai
-
-1. Go to Claude.ai → Settings → Integrations
-2. Click "Add More"
-3. Enter your server URL: `https://abc123.ngrok.io/mcp`
-4. Click "Connect"
-5. Authorize the connection
-
-### 5. Use in Conversations
-
-- Click the tools icon (🔧) in the chat
-- Select "Basic Memory"
-- Try commands like:
-  - "Create a note about our meeting"
-  - "Search for project ideas"
-  - "Show recent notes"
-
-## Production Setup
-
-### 1. Deploy with Supabase Auth
-
-```bash
-# .env file
-FASTMCP_AUTH_ENABLED=true
-FASTMCP_AUTH_PROVIDER=supabase
-SUPABASE_URL=https://your-project.supabase.co
-SUPABASE_ANON_KEY=your-anon-key
-SUPABASE_SERVICE_KEY=your-service-key
-```
-
-### 2. Deploy to Cloud
-
-Options for deployment:
-
-#### Vercel
-```json
-// vercel.json
-{
-  "functions": {
-    "api/mcp.py": {
-      "runtime": "python3.9"
-    }
-  }
-}
-```
-
-#### Railway
-```bash
-# Install Railway CLI
-brew install railway
-
-# Deploy
-railway init
-railway up
-```
-
-#### Docker
-```dockerfile
-FROM python:3.12
-WORKDIR /app
-COPY . .
-RUN pip install -e .
-CMD ["basic-memory", "mcp", "--transport", "streamable-http"]
-```
-
-### 3. Configure for Organization
-
-For Claude.ai Enterprise:
-
-1. **Admin Setup**:
-   - Go to Organizational Settings
-   - Navigate to Integrations
-   - Add MCP server URL for all users
-   - Configure allowed scopes
-
-2. **User Permissions**:
-   - Users connect individually
-   - Each user has their own auth token
-   - Scopes determine access level
-
-## Security Best Practices
-
-### 1. Use HTTPS
-- Required for OAuth
-- Encrypt all data in transit
-- Use proper SSL certificates
-
-### 2. Implement Scopes
-```bash
-# Configure required scopes
-FASTMCP_AUTH_REQUIRED_SCOPES=read,write
-
-# User-specific scopes
-read: Can search and read notes
-write: Can create and update notes
-admin: Can manage all data
-```
-
-### 3. Token Security
-- Short-lived access tokens (1 hour)
-- Refresh token rotation
-- Secure token storage
-
-### 4. Rate Limiting
-```python
-# In your MCP server
-from fastapi import HTTPException
-from slowapi import Limiter
-
-limiter = Limiter(key_func=get_remote_address)
-
-@app.get("/mcp")
-@limiter.limit("100/minute")
-async def mcp_endpoint():
-    # Handle MCP requests
-```
-
-## Advanced Features
-
-### 1. Custom Tools
-
-Create specialized tools for Claude:
-
-```python
-@mcp.tool()
-async def analyze_notes(topic: str) -> str:
-    """Analyze all notes on a specific topic."""
-    # Search and analyze implementation
-    return analysis
-```
-
-### 2. Context Preservation
-
-Use memory:// URLs to maintain context:
-
-```python
-@mcp.tool()
-async def continue_conversation(memory_url: str) -> str:
-    """Continue from a previous conversation."""
-    context = await build_context(memory_url)
-    return context
-```
-
-### 3. Multi-User Support
-
-With Supabase, each user has isolated data:
-
-```sql
--- Row-level security
-CREATE POLICY "Users see own notes"
-ON notes FOR SELECT
-USING (auth.uid() = user_id);
-```
-
-## Troubleshooting
-
-### Connection Issues
-
-1. **"Failed to connect"**
-   - Verify server is running
-   - Check HTTPS is working
-   - Confirm OAuth is enabled
-
-2. **"Authorization failed"**
-   - Check client credentials
-   - Verify redirect URLs
-   - Review OAuth logs
-
-3. **"No tools available"**
-   - Ensure MCP tools are registered
-   - Check required scopes
-   - Verify transport type
-
-### Debug Mode
-
-Enable detailed logging:
-
-```bash
-# Server side
-export FASTMCP_LOG_LEVEL=DEBUG
-export LOGURU_LEVEL=DEBUG
-
-# Check logs
-tail -f logs/mcp.log
-```
-
-### Test Connection
-
-```bash
-# Test OAuth flow
-curl https://your-server.com/mcp/.well-known/oauth-authorization-server
-
-# Should return OAuth metadata
-{
-  "issuer": "https://your-server.com",
-  "authorization_endpoint": "https://your-server.com/auth/authorize",
-  ...
-}
-```
-
-## Best Practices
-
-1. **Regular Backups**
-   - Export your knowledge base
-   - Use version control
-   - Multiple storage locations
-
-2. **Access Control**
-   - Principle of least privilege
-   - Regular token rotation
-   - Audit access logs
-
-3. **Performance**
-   - Index frequently searched fields
-   - Optimize large knowledge bases
-   - Use caching where appropriate
-
-4. **User Experience**
-   - Clear tool descriptions
-   - Helpful error messages
-   - Quick response times
-
-## Examples
-
-### Creating Notes
-
-```
-User: Create a note about the meeting with the product team
-
-Claude: I'll create a note about your meeting with the product team.
-
-[Uses write_note tool]
-
-Note created: "Meeting with Product Team - 2024-01-15"
-Location: Work/Meetings/
-
-I've documented the meeting notes. The note includes the date, attendees, and key discussion points.
-```
-
-### Searching Knowledge
-
-```
-User: What did we discuss about the API redesign?
-
-Claude: Let me search for information about the API redesign.
-
-[Uses search_notes tool]
-
-I found 3 relevant notes about the API redesign:
-
-1. "API Redesign Proposal" (2024-01-10)
-   - RESTful architecture
-   - Version 2.0 specifications
-   - Migration timeline
-
-2. "Technical Review: API Changes" (2024-01-12)
-   - Breaking changes documented
-   - Backwards compatibility plan
-
-3. "Meeting: API Implementation" (2024-01-14)
-   - Team assignments
-   - Q1 deliverables
-```
-
-## Next Steps
-
-1. Set up production deployment
-2. Configure organizational access
-3. Create custom tools for your workflow
-4. Implement advanced security features
-5. Monitor usage and performance
-
-## Resources
-
-- [Basic Memory Documentation](../README.md)
-- [OAuth Setup Guide](OAuth%20Authentication.md)
-- [MCP Specification](https://modelcontextprotocol.io)
-- [Claude.ai Help Center](https://support.anthropic.com)

docs/Docker.md

@@ -0,0 +1,365 @@
+# Docker Setup Guide
+
+Basic Memory can be run in Docker containers to provide a consistent, isolated environment for your knowledge management
+system. This is particularly useful for integrating with existing Dockerized MCP servers or for deployment scenarios.
+
+## Quick Start
+
+### Option 1: Using Pre-built Images (Recommended)
+
+Basic Memory provides pre-built Docker images on GitHub Container Registry that are automatically updated with each release.
+
+1. **Use the official image directly:**
+   ```bash
+   docker run -d \
+     --name basic-memory-server \
+     -p 8000:8000 \
+     -v /path/to/your/obsidian-vault:/app/data:rw \
+     -v basic-memory-config:/app/.basic-memory:rw \
+     ghcr.io/basicmachines-co/basic-memory:latest
+   ```
+
+2. **Or use Docker Compose with the pre-built image:**
+   ```yaml
+   version: '3.8'
+   services:
+     basic-memory:
+       image: ghcr.io/basicmachines-co/basic-memory:latest
+       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
+   ```
+
+### Option 2: Using Docker Compose (Building Locally)
+
+1. **Clone the repository:**
+   ```bash
+   git clone https://github.com/basicmachines-co/basic-memory.git
+   cd basic-memory
+   ```
+
+2. **Update the docker-compose.yml:**
+   Edit the volume mount to point to your Obsidian vault:
+   ```yaml
+   volumes:
+     # Change './obsidian-vault' to your actual directory path
+     - /path/to/your/obsidian-vault:/app/data:rw
+   ```
+
+3. **Start the container:**
+   ```bash
+   docker-compose up -d
+   ```
+
+### Option 3: Using Docker CLI
+
+```bash
+# Build the image
+docker build -t basic-memory .
+
+# Run with volume mounting
+docker run -d \
+  --name basic-memory-server \
+  -v /path/to/your/obsidian-vault:/app/data:rw \
+  -v basic-memory-config:/app/.basic-memory:rw \
+  -e BASIC_MEMORY_DEFAULT_PROJECT=main \
+  basic-memory
+```
+
+## Configuration
+
+### Volume Mounts
+
+Basic Memory requires several volume mounts for proper operation:
+
+1. **Knowledge Directory** (Required):
+   ```yaml
+   - /path/to/your/obsidian-vault:/app/data:rw
+   ```
+   Mount your Obsidian vault or knowledge base directory.
+
+2. **Configuration and Database** (Recommended):
+   ```yaml
+   - basic-memory-config:/app/.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.
+
+3. **Multiple Projects** (Optional):
+   ```yaml
+   - /path/to/project1:/app/data/project1:rw
+   - /path/to/project2:/app/data/project2:rw
+   ```
+
+You can edit the basic-memory config.json file located in the /app/.basic-memory/config.json
+
+## CLI Commands via Docker
+
+You can run Basic Memory CLI commands inside the container using `docker exec`:
+
+### Basic Commands
+
+```bash
+# Check status
+docker exec basic-memory-server basic-memory status
+
+# Sync files
+docker exec basic-memory-server basic-memory sync
+
+# Show help
+docker exec basic-memory-server basic-memory --help
+```
+
+### Managing Projects with Volume Mounts
+
+When using Docker volumes, you'll need to configure projects to point to your mounted directories:
+
+1. **Check current configuration:**
+   ```bash
+   docker exec basic-memory-server cat /app/.basic-memory/config.json
+   ```
+
+2. **Add a project for your mounted volume:**
+   ```bash
+   # If you mounted /path/to/your/vault to /app/data
+   docker exec basic-memory-server basic-memory project create my-vault /app/data
+   
+   # Set it as default
+   docker exec basic-memory-server basic-memory project set-default my-vault
+   ```
+
+3. **Sync the new project:**
+   ```bash
+   docker exec basic-memory-server basic-memory sync
+   ```
+
+### Example: Setting up an Obsidian Vault
+
+If you mounted your Obsidian vault like this in docker-compose.yml:
+```yaml
+volumes:
+  - /Users/yourname/Documents/ObsidianVault:/app/data:rw
+```
+
+Then configure it:
+```bash
+# Create project pointing to mounted vault
+docker exec basic-memory-server basic-memory project create obsidian /app/data
+
+# Set as default
+docker exec basic-memory-server basic-memory project set-default obsidian
+
+# Sync to index all files
+docker exec basic-memory-server basic-memory sync
+```
+
+### Environment Variables
+
+Configure Basic Memory using environment variables:
+
+```yaml
+environment:
+
+  # Default project
+  - BASIC_MEMORY_DEFAULT_PROJECT=main
+
+  # Enable real-time sync
+  - BASIC_MEMORY_SYNC_CHANGES=true
+
+  # Logging level
+  - BASIC_MEMORY_LOG_LEVEL=INFO
+
+  # Sync delay in milliseconds
+  - BASIC_MEMORY_SYNC_DELAY=1000
+```
+
+## File Permissions
+
+### 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:
+
+```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 .
+
+# Or use docker-compose with build args
+```
+
+**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:
+
+1. Open Docker Desktop
+2. Go to Settings → Resources → File Sharing
+3. Add your knowledge directory path
+4. Apply & Restart
+
+## Troubleshooting
+
+### Common Issues
+
+1. **File Watching Not Working:**
+    - Ensure volume mounts are read-write (`:rw`)
+    - Check directory permissions
+    - On Linux, may need to increase inotify limits:
+      ```bash
+      echo fs.inotify.max_user_watches=524288 | sudo tee -a /etc/sysctl.conf
+      sudo sysctl -p
+      ```
+
+2. **Configuration Not Persisting:**
+    - Use named volumes for `/app/.basic-memory`
+    - Check volume mount permissions
+
+3. **Network Connectivity:**
+    - For HTTP transport, ensure port 8000 is exposed
+    - Check firewall settings
+
+### Debug Mode
+
+Run with debug logging:
+
+```yaml
+environment:
+  - BASIC_MEMORY_LOG_LEVEL=DEBUG
+```
+
+View logs:
+
+```bash
+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.
+
+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.
+
+3. **Network Security:**
+   If using HTTP transport, consider using reverse proxy with SSL/TLS and authentication if the endpoint is available on
+   a network.
+
+4. **IMPORTANT:** The HTTP endpoints have no authorization. They should not be exposed on a public network.  
+
+## Integration Examples
+
+### Claude Desktop with Docker
+
+The recommended way to connect Claude Desktop to the containerized Basic Memory is using `mcp-proxy`, which converts the HTTP transport to STDIO that Claude Desktop expects:
+
+1. **Start the Docker container:**
+   ```bash
+   docker-compose up -d
+   ```
+
+2. **Configure Claude Desktop** to use mcp-proxy:
+   ```json
+   {
+     "mcpServers": {
+       "basic-memory": {
+         "command": "uvx",
+         "args": [
+           "mcp-proxy",
+           "http://localhost:8000/mcp"
+         ]
+       }
+     }
+   }
+   ```
+
+
+## Support
+
+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`
+
+For general Basic Memory support, see the main [README](../README.md)
+and [documentation](https://memory.basicmachines.co/).
+
+## GitHub Container Registry Images
+
+### Available Images
+
+Pre-built Docker images are available on GitHub Container Registry at [`ghcr.io/basicmachines-co/basic-memory`](https://github.com/basicmachines-co/basic-memory/pkgs/container/basic-memory).
+
+**Supported architectures:**
+- `linux/amd64` (Intel/AMD x64)
+- `linux/arm64` (ARM64, including Apple Silicon)
+
+**Available tags:**
+- `latest` - Latest stable release
+- `v0.13.8`, `v0.13.7`, etc. - Specific version tags
+- `v0.13`, `v0.12`, etc. - Major.minor tags
+
+### Automated Builds
+
+Docker images are automatically built and published when new releases are tagged:
+
+1. **Release Process:** When a git tag matching `v*` (e.g., `v0.13.8`) is pushed, the CI workflow automatically:
+   - Builds multi-platform Docker images
+   - Pushes to GitHub Container Registry with appropriate tags
+   - Uses native GitHub integration for seamless publishing
+
+2. **CI/CD Pipeline:** The Docker workflow includes:
+   - Multi-platform builds (AMD64 and ARM64)
+   - Layer caching for faster builds
+   - Automatic tagging with semantic versioning
+   - Security scanning and optimization
+
+### Setup Requirements (For Maintainers)
+
+GitHub Container Registry integration is automatic for this repository:
+
+1. **No external setup required** - GHCR is natively integrated with GitHub
+2. **Automatic permissions** - Uses `GITHUB_TOKEN` with `packages: write` permission
+3. **Public by default** - Images are automatically public for public repositories
+
+The Docker CI workflow (`.github/workflows/docker.yml`) handles everything automatically when version tags are pushed.

docs/Getting Started with Basic Memory.md

@@ -1,355 +0,0 @@
----
-title: Getting Started with Basic Memory
-type: note
-permalink: docs/getting-started
----
-
-# Getting Started with Basic Memory
-
-This guide will help you install Basic Memory, configure it with Claude Desktop, and create your first knowledge notes
-through conversations.
-
-Basic Memory uses the [Model Context Protocol](https://modelcontextprotocol.io/introduction) (MCP) to connect with LLMs.
-It can be used with any service that supports the MCP, but Claude Desktop works especially well.
-
-## Installation
-
-### Prerequisites
-
-The easiest way to install basic memory is via `uv`. See the [uv installation guide](https://docs.astral.sh/uv/getting-started/installation/).
-
-### 1. Install Basic Memory
-
-**v0.13.0 offers multiple installation options:**
-
-```bash
-# Stable release (recommended)
-uv tool install basic-memory
-# or: pip install basic-memory
-
-# Beta releases (new features, testing)
-pip install basic-memory --pre
-
-# Development builds (latest changes)
-pip install basic-memory --pre --force-reinstall
-```
-
-**Version Information:**
-- **Stable**: Latest tested release (e.g., `0.13.0`)
-- **Beta**: Pre-release versions (e.g., `0.13.0b1`) 
-- **Development**: Auto-published from git commits (e.g., `0.12.4.dev26+468a22f`)
-
-> **Important**: You need to install Basic Memory using one of the commands above to use the command line tools.
-
-Using `uv tool install` will install the basic-memory package in a standalone virtual environment. See the [UV docs](https://docs.astral.sh/uv/concepts/tools/) for more info.
-
-### 2. Configure Claude Desktop
-
-Edit your Claude Desktop config, located at `~/Library/Application Support/Claude/claude_desktop_config.json`:
-
-```json
-{
-  "mcpServers": {
-    "basic-memory": {
-      "command": "uvx",
-      "args": [
-        "basic-memory",
-        "mcp"
-      ]
-    }
-  }
-}
-```
-
-**Restart Claude Desktop**. You should see Basic Memory tools available in the "tools" menu in Claude Desktop (the little hammer icon in the bottom-right corner of the chat interface). Click it to view available tools.
-#### Fix Path to uv
-
-If you get an error that says `ENOENT` , this most likely means Claude Desktop could not find your `uv` installation. Make sure that you have `uv` installed per the instructions above, then:
-
-**Step 1: Find the absolute path to uvx**
-
-Open Terminal and run:
-
-```bash
-which uvx
-```
-
-This will show you the full path (e.g., `/Users/yourusername/.cargo/bin/uvx`).
-
-**Step 2: Edit Claude Desktop Configuration**
-
-Edit the Claude Desktop config:
-
-```json
-{
-  "mcpServers": {
-    "basic-memory": {
-      "command": "/absolute/path/to/uvx",
-      "args": [
-        "basic-memory",
-        "mcp"
-      ]
-    }
-  }
-}
-```
-
-Replace `/absolute/path/to/uvx` with the actual path you found in Step 1.
-
-**Step 3: Restart Claude Desktop**
-
-Close and reopen Claude Desktop for the changes to take effect.
-
-### 3. Sync changes in real time
-
-> **Note**: The service will sync changes from your project directory in real time so they available for the AI assistant. 
-
-To disable realtime sync, you can update the config. See [[CLI Reference#sync]].
-### 4. Staying Updated
-
-To update Basic Memory when new versions are released:
-
-```bash
-# Update stable release
-uv tool upgrade basic-memory 
-# or: pip install --upgrade basic-memory
-
-# Update to latest beta (v0.13.0)
-pip install --upgrade basic-memory --pre
-
-# Get latest development build
-pip install --upgrade basic-memory --pre --force-reinstall
-```
-
-**v0.13.0 Update Benefits:**
-- **Fluid project switching** during conversations
-- **Advanced note editing** capabilities
-- **Smart file management** with move operations
-- **Enhanced search** with frontmatter tag support
-
-> **Note**: After updating, restart Claude Desktop for changes to take effect. No sync restart needed in v0.13.0.
-
-### 5. Multi-Project Setup (Enhanced in v0.13.0)
-
-By default, Basic Memory creates a project in `~/basic-memory`. v0.13.0 introduces **fluid project management** - switch between projects instantly during conversations.
-
-```  
-# Create a new project
-basic-memory project create work ~/work-basic-memory  
-  
-# Set the default project  
-basic-memory project set-default work  
-
-# List all projects with status
-basic-memory project list
-
-# Get detailed project information
-basic-memory project info
-```
-
-**New in v0.13.0:**
-- **Instant switching**: Change projects during conversations without restart
-- **Unified database**: All projects in single `~/.basic-memory/memory.db`
-- **Better performance**: Optimized queries and reduced file I/O
-- **Session context**: Maintains active project throughout conversations
-
-## Troubleshooting Installation
-
-### Common Issues
-
-#### Claude Says "No Basic Memory Tools Available"
-
-If Claude cannot find Basic Memory tools:
-
-1. **Check absolute paths**: Ensure you're using complete absolute paths to uvx in the Claude Desktop configuration
-2. **Verify installation**: Run `basic-memory --version` in Terminal to confirm Basic Memory is installed
-3. **Restart applications**: Restart both Terminal and Claude Desktop after making configuration changes
-4. **Check sync status**: You can view the sync status by running `basic-memory status
-.
-#### Permission Issues
-
-If you encounter permission errors:
-
-1. Check that Basic Memory has access to create files in your home directory
-2. Ensure Claude Desktop has permission to execute the uvx command
-
-## Creating Your First Knowledge Note
-
-1. **Open Claude Desktop** and start a new conversation.
-
-2. **Have a natural conversation** about any topic:
-   ```
-   You: "Let's talk about coffee brewing methods I've been experimenting with."
-   Claude: "I'd be happy to discuss coffee brewing methods..."
-   You: "I've found that pour over gives more flavor clarity than French press..."
-   ```
-
-3. **Ask Claude to create a note**:
-   ```
-   You: "Could you create a note summarizing what we've discussed about coffee brewing?"
-   ```
-
-4. **Confirm note creation**:
-   Claude will confirm when the note has been created and where it's stored.
-
-5. **View the created file** in your `~/basic-memory` directory using any text editor or Obsidian.
-   The file structure will look similar to:
-   ```markdown
-   ---
-   title: Coffee Brewing Methods
-   permalink: coffee-brewing-methods
-   tags: [coffee, brewing, equipment]  # v0.13.0: Now searchable!
-   ---
-   
-   # Coffee Brewing Methods
-   
-   ## Observations
-   - [method] Pour over provides more clarity...
-   - [technique] Water temperature at 205°F...
-   
-   ## Relations
-   - relates_to [[Other Coffee Topics]]
-   ```
-
-**v0.13.0 Improvements:**
-- **Real-time sync**: Changes appear immediately, no background sync needed
-- **Searchable tags**: Frontmatter tags are now indexed for search
-- **Better file organization**: Enhanced file management capabilities
-
-## Using Special Prompts
-
-Basic Memory includes special prompts that help you start conversations with context from your knowledge base:
-
-### Continue Conversation
-
-To resume a previous topic:
-
-```
-You: "Let's continue our conversation about coffee brewing."
-```
-
-This prompt triggers Claude to:
-
-1. Search your knowledge base for relevant content about coffee brewing
-2. Build context from these documents
-3. Resume the conversation with full awareness of previous discussions
-
-### Recent Activity
-
-To see what you've been working on:
-
-```
-You: "What have we been discussing recently?"
-```
-
-This prompt causes Claude to:
-
-1. Retrieve documents modified in the recent past
-2. Summarize the topics and main points
-3. Offer to continue any of those discussions
-
-### Search
-
-To find specific information:
-
-```
-You: "Find information about pour over coffee methods."
-```
-
-Claude will:
-
-1. Search your knowledge base for relevant documents
-2. Summarize the key findings
-3. Offer to explore specific documents in more detail
-
-See [[User Guide#Using Special Prompts]] for further information.
-
-## Using Your Knowledge Base
-
-### Referencing Knowledge
-
-In future conversations, reference your existing knowledge:
-
-```
-You: "What water temperature did we decide was optimal for coffee brewing?"
-```
-
-Or directly reference notes using memory:// URLs:
-
-```
-You: "Take a look at memory://coffee-brewing-methods and let's discuss how to improve my technique."
-```
-
-### Building On Previous Knowledge (Enhanced in v0.13.0)
-
-Basic Memory enables continuous knowledge building:
-
-1. **Reference previous discussions** in new conversations
-2. **Edit notes incrementally** without rewriting entire documents
-3. **Move and organize notes** as your knowledge base grows
-4. **Switch between projects** instantly during conversations
-5. **Search by tags** to find related content quickly
-6. **Create connections** between related topics
-7. **Follow relationships** to build comprehensive context
-
-### v0.13.0 Workflow Examples
-
-**Incremental Editing:**
-```
-You: "Add a section about espresso to my coffee brewing notes"
-Claude: [Uses edit_note to append new section]
-```
-
-**File Organization:**
-```
-You: "Move my old meeting notes to an archive folder"
-Claude: [Uses move_note with database consistency]
-```
-
-**Project Switching:**
-```
-You: "Switch to my work project and show recent activity"
-Claude: [Switches projects and shows work-specific content]
-```
-
-## Importing Existing Conversations
-
-Import your existing AI conversations:
-
-```bash
-# From Claude
-basic-memory import claude conversations
-
-# From ChatGPT
-basic-memory import chatgpt
-```
-
-After importing, changes sync automatically in real-time. You can see project statistics by running `basic-memory project info`.
-
-## Quick Tips
-
-### General Usage
-- Basic Memory syncs changes in real-time (no manual sync needed)
-- Use special prompts (Continue Conversation, Recent Activity, Search) to start contextual discussions
-- Build connections between notes for a richer knowledge graph
-- Use direct `memory://` URLs with permalinks for precise context
-- Review and edit AI-generated notes for accuracy
-
-### v0.13.0 Features
-- **Switch projects instantly**: "Switch to my work project" - no restart needed
-- **Edit notes incrementally**: "Add a section about..." instead of rewriting
-- **Organize with moves**: "Move this to my archive folder" with database consistency
-- **Search by tags**: Frontmatter tags are now searchable
-- **Try beta builds**: `pip install basic-memory --pre` for latest features
-
-## Next Steps
-
-After getting started, explore these areas:
-
-1. **Read the [[User Guide]]** for comprehensive usage instructions
-2. **Understand the [[Knowledge Format]]** to learn how knowledge is structured
-3. **Set up [[Obsidian Integration]]** for visual knowledge navigation
-4. **Learn about [[Canvas]]** visualizations for mapping concepts
-5. **Review the [[CLI Reference]]** for command line tools
-6. **Explore [[OAuth Authentication Guide]]** for secure remote access (v0.13.0)
-7. **Set up multiple projects** for different knowledge areas (v0.13.0)

docs/Knowledge Format.md

@@ -1,207 +0,0 @@
----
-title: Knowledge Format
-type: note
-permalink: docs/knowledge-format
-tags:
-- architecture
-- patterns
-- knowledge
-- design
----
-
-# Knowledge Format
-
-Basic Memory uses standard Markdown with simple semantic patterns to create a knowledge graph. This document details the file structure and patterns used to organize knowledge.
-
-## File-First Architecture
-
-All knowledge in Basic Memory is stored in plain text Markdown files:
-
-- Files are the source of truth for all knowledge
-- Changes to files automatically update the knowledge graph
-- You maintain complete ownership and control
-- Files work with git and other version control systems
-- Knowledge persists independently of any AI conversation
-
-## Core Document Structure
-
-Every document uses this basic structure:
-
-```markdown
----
-title: Document Title
-type: note
-tags: [tag1, tag2]
-permalink: custom-path
----
-
-# Document Title
-
-Regular markdown content...
-
-## Observations
-- [category] Content with #tags (optional context)
-
-## Relations
-- relation_type [[Other Document]] (optional context)
-```
-
-### Frontmatter
-
-The YAML frontmatter at the top of each file defines essential metadata:
-
-```yaml
----
-title: Document Title    # Used for linking and references
-type: note               # Document type
-tags: [tag1, tag2]       # For organization and searching
-permalink: custom-link   # Optional custom URL path
----
-```
-
-The title is particularly important as it's used to create links between documents.
-
-### Observations
-
-Observations are facts or statements about a topic:
-
-```markdown
-## Observations
-- [tech] Uses SQLite for storage #database
-- [design] Follows local-first architecture #architecture
-- [decision] Selected bcrypt for passwords #security (Based on audit)
-```
-
-Each observation contains:
-- **Category** in [brackets] - classifies the information type
-- **Content text** - the main information
-- Optional **#tags** - additional categorization
-- Optional **(context)** - supporting details
-
-Common categories include:
-- `[tech]`: Technical details
-- `[design]`: Architecture decisions
-- `[feature]`: User capabilities
-- `[decision]`: Choices that were made
-- `[principle]`: Fundamental concepts
-- `[method]`: Approaches or techniques
-- `[preference]`: Personal opinions
-
-### Relations
-
-Relations connect documents to form the knowledge graph:
-
-```markdown
-## Relations
-- implements [[Search Design]]
-- depends_on [[Database Schema]]
-- relates_to [[User Interface]]
-```
-
-You can also create inline references:
-```markdown
-This builds on [[Core Design]] and uses [[Utility Functions]].
-```
-
-Common relation types include:
-- `implements`: Implementation of a specification
-- `depends_on`: Required dependency
-- `relates_to`: General connection
-- `inspired_by`: Source of ideas
-- `extends`: Enhancement
-- `part_of`: Component relationship
-- `contains`: Hierarchical relationship
-- `pairs_with`: Complementary relationship
-
-## Knowledge Graph
-
-Basic Memory automatically builds a knowledge graph from your document connections:
-
-- Each document becomes a node in the graph
-- Relations create edges between nodes
-- Relation types add semantic meaning to connections
-- Forward references can link to documents that don't exist yet
-
-This graph enables rich context building and navigation across your knowledge base.
-
-## Permalinks and memory:// URLs
-
-Every document in Basic Memory has a unique permalink that serves as its stable identifier:
-
-### How Permalinks Work
-
-- **Automatically assigned**: The system generates a permalink for each document
-- **Based on title**: By default, derived from the document title
-- **Always unique**: If conflicts exist, the system adds a suffix to ensure uniqueness
-- **Stable reference**: Remains the same even if the file moves in the directory structure
-- **Used in memory:// URLs**: Forms the basis of the memory:// addressing scheme
-
-You can specify a custom permalink in the frontmatter:
-```yaml
----
-title: Authentication Approaches
-permalink: auth-approaches-2024
----
-```
-
-If not specified, one will be generated automatically from the title, if the note has has a frontmatter section. 
-
-By default a  notes' permalink value will not change if the file is moved. It's a **stable** identifier :). But if you'd rather permalinks are always updated when a file moves, you can set the config setting in the global config. 
-
-The config file for Basic Memory is in the home directory under `.basic-memory/config.json`.
-
-To change the behavior, set the following value:
-```
- ~/.basic-memory/config.json 
-{
-  "update_permalinks_on_move": true
-}
-```
-
-### Using memory:// URLs
-
-The memory:// URL scheme provides a reliable way to reference knowledge:
-
-```
-memory://auth-approaches-2024        # Direct access by permalink
-memory://Authentication Approaches   # Access by title (automatically resolves)
-memory://project/auth-approaches     # Access by path
-```
-
-Memory URLs support pattern matching for more powerful queries:
-
-```
-memory://auth*                      # All documents with permalinks starting with "auth"
-memory://*/approaches               # All documents with permalinks ending with "approaches"
-memory://project/*/requirements     # All requirements documents in the project folder
-memory://docs/search/implements/*   # Follow all implements relations from search docs
-```
-
-This addressing scheme ensures content remains accessible even as your knowledge base evolves and files are reorganized.
-
-## File Organization
-
-Organize files in any structure that suits your needs:
-
-```
-docs/
-  architecture/
-    design.md
-    patterns.md
-  features/
-    search.md
-    auth.md
-```
-
-You can:
-- Group by topic in folders
-- Use a flat structure with descriptive filenames
-- Tag files for easier discovery
-- Add custom metadata in frontmatter
-
-The system will build the semantic knowledge graph regardless of how you organize your files.
-
-## Relations
-- implemented_by [[User Guide]] (How to work with this format)
-- relates_to [[Getting Started with Basic Memory]] (Setup instructions)
-- explained_in [[Introduction to Basic Memory]] (Overview of the system)

docs/NOTE-FORMAT.md

@@ -0,0 +1,494 @@
+# 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/OAuth Authentication Guide.md

@@ -1,259 +0,0 @@
-# OAuth Authentication Guide
-
-Basic Memory MCP server supports OAuth 2.1 authentication for secure access control. This guide covers setup, testing, and production deployment.
-
-## Quick Start
-
-### 1. Enable OAuth
-
-```bash
-# Set environment variable
-export FASTMCP_AUTH_ENABLED=true
-
-# Or use .env file
-echo "FASTMCP_AUTH_ENABLED=true" >> .env
-```
-
-### 2. Start the Server
-
-```bash
-basic-memory mcp --transport streamable-http
-```
-
-### 3. Test with MCP Inspector
-
-Since the basic auth provider uses in-memory storage with per-instance secret keys, you'll need to use a consistent approach:
-
-#### Option A: Use Environment Variable for Secret Key
-
-```bash
-# Set a fixed secret key for testing
-export FASTMCP_AUTH_SECRET_KEY="your-test-secret-key"
-
-# Start the server
-FASTMCP_AUTH_ENABLED=true basic-memory mcp --transport streamable-http
-
-# In another terminal, register a client
-basic-memory auth register-client --client-id=test-client
-
-# Get a token using the same secret key
-basic-memory auth test-auth
-```
-
-#### Option B: Use the Built-in Test Endpoint
-
-```bash
-# Start server with OAuth
-FASTMCP_AUTH_ENABLED=true basic-memory mcp --transport streamable-http
-
-# Register a client and get token in one step
-curl -X POST http://localhost:8000/register \
-  -H "Content-Type: application/json" \
-  -d '{"client_metadata": {"client_name": "Test Client"}}'
-
-# Use the returned client_id and client_secret
-curl -X POST http://localhost:8000/token \
-  -H "Content-Type: application/x-www-form-urlencoded" \
-  -d "grant_type=client_credentials&client_id=YOUR_CLIENT_ID&client_secret=YOUR_CLIENT_SECRET"
-```
-
-### 4. Configure MCP Inspector
-
-1. Open MCP Inspector
-2. Configure:
-   - Server URL: `http://localhost:8000/mcp/` (note the trailing slash!)
-   - Transport: `streamable-http`
-   - Custom Headers:
-     ```
-     Authorization: Bearer YOUR_ACCESS_TOKEN
-     Accept: application/json, text/event-stream
-     ```
-
-## OAuth Endpoints
-
-The server provides these OAuth endpoints automatically:
-
-- `GET /authorize` - Authorization endpoint
-- `POST /token` - Token exchange endpoint
-- `GET /.well-known/oauth-authorization-server` - OAuth metadata
-- `POST /register` - Client registration (if enabled)
-- `POST /revoke` - Token revocation (if enabled)
-
-## OAuth Flow
-
-### Standard Authorization Code Flow
-
-1. **Get Authorization Code**:
-   ```bash
-   curl "http://localhost:8000/authorize?client_id=YOUR_CLIENT_ID&redirect_uri=http://localhost:8000/callback&response_type=code&code_challenge=YOUR_CHALLENGE&code_challenge_method=S256"
-   ```
-
-2. **Exchange Code for Token**:
-   ```bash
-   curl -X POST http://localhost:8000/token \
-     -H "Content-Type: application/x-www-form-urlencoded" \
-     -d "grant_type=authorization_code&code=AUTH_CODE&client_id=CLIENT_ID&client_secret=CLIENT_SECRET&code_verifier=YOUR_VERIFIER"
-   ```
-
-3. **Use Access Token**:
-   ```bash
-   curl http://localhost:8000/mcp \
-     -H "Authorization: Bearer ACCESS_TOKEN"
-   ```
-
-## Production Deployment
-
-### Using Supabase Auth
-
-For production, use Supabase for persistent auth storage:
-
-```bash
-# Configure environment
-FASTMCP_AUTH_ENABLED=true
-FASTMCP_AUTH_PROVIDER=supabase
-SUPABASE_URL=https://your-project.supabase.co
-SUPABASE_ANON_KEY=your-anon-key
-SUPABASE_SERVICE_KEY=your-service-key
-
-# Start server
-basic-memory mcp --transport streamable-http --host 0.0.0.0
-```
-
-### Security Requirements
-
-1. **HTTPS Required**: OAuth requires HTTPS in production (localhost exception for testing)
-2. **PKCE Support**: Claude.ai requires PKCE for authorization
-3. **Token Expiration**: Access tokens expire after 1 hour
-4. **Scopes**: Supported scopes are `read`, `write`, and `admin`
-
-## Connecting from Claude.ai
-
-1. **Deploy with HTTPS**:
-   ```bash
-   # Use ngrok for testing
-   ngrok http 8000
-   
-   # Or deploy to cloud provider
-   ```
-
-2. **Configure in Claude.ai**:
-   - Go to Settings → Integrations
-   - Click "Add More"
-   - Enter: `https://your-server.com/mcp`
-   - Click "Connect"
-   - Authorize in the popup window
-
-## Debugging
-
-### Common Issues
-
-1. **401 Unauthorized**: 
-   - Check token is valid and not expired
-   - Verify secret key consistency
-   - Ensure bearer token format: `Authorization: Bearer TOKEN`
-
-2. **404 on Auth Endpoints**:
-   - Endpoints are at root, not under `/auth`
-   - Use `/authorize` not `/auth/authorize`
-
-3. **Token Validation Fails**:
-   - Basic provider uses in-memory storage
-   - Tokens don't persist across server restarts
-   - Use same secret key for testing
-
-### Debug Commands
-
-```bash
-# Check OAuth metadata
-curl http://localhost:8000/.well-known/oauth-authorization-server
-
-# Enable debug logging
-export FASTMCP_LOG_LEVEL=DEBUG
-
-# Test token directly
-curl http://localhost:8000/mcp \
-  -H "Authorization: Bearer YOUR_TOKEN" \
-  -v
-```
-
-## Provider Options
-
-- **basic**: In-memory storage (development only)
-- **supabase**: Recommended for production
-- **github**: GitHub OAuth integration
-- **google**: Google OAuth integration
-
-## Example Test Script
-
-```python
-import httpx
-import asyncio
-from urllib.parse import urlparse, parse_qs
-
-async def test_oauth_flow():
-    """Test the full OAuth flow"""
-    client_id = "test-client"
-    client_secret = "test-secret"
-    
-    async with httpx.AsyncClient() as client:
-        # 1. Get authorization code
-        auth_response = await client.get(
-            "http://localhost:8000/authorize",
-            params={
-                "client_id": client_id,
-                "redirect_uri": "http://localhost:8000/callback",
-                "response_type": "code",
-                "code_challenge": "test-challenge",
-                "code_challenge_method": "S256",
-                "state": "test-state"
-            }
-        )
-        
-        # Extract code from redirect URL
-        redirect_url = auth_response.headers.get("Location")
-        parsed = urlparse(redirect_url)
-        code = parse_qs(parsed.query)["code"][0]
-        
-        # 2. Exchange for token
-        token_response = await client.post(
-            "http://localhost:8000/token",
-            data={
-                "grant_type": "authorization_code",
-                "code": code,
-                "client_id": client_id,
-                "client_secret": client_secret,
-                "code_verifier": "test-verifier",
-                "redirect_uri": "http://localhost:8000/callback"
-            }
-        )
-        
-        tokens = token_response.json()
-        print(f"Access token: {tokens['access_token']}")
-        
-        # 3. Test MCP endpoint
-        mcp_response = await client.post(
-            "http://localhost:8000/mcp",
-            headers={"Authorization": f"Bearer {tokens['access_token']}"},
-            json={"method": "initialize", "params": {}}
-        )
-        
-        print(f"MCP Response: {mcp_response.status_code}")
-
-asyncio.run(test_oauth_flow())
-```
-
-## Environment Variables
-
-| Variable | Description | Default |
-|----------|-------------|---------|
-| `FASTMCP_AUTH_ENABLED` | Enable OAuth authentication | `false` |
-| `FASTMCP_AUTH_PROVIDER` | OAuth provider type | `basic` |
-| `FASTMCP_AUTH_SECRET_KEY` | JWT signing key (basic provider) | Random |
-| `FASTMCP_AUTH_ISSUER_URL` | OAuth issuer URL | `http://localhost:8000` |
-| `FASTMCP_AUTH_REQUIRED_SCOPES` | Required scopes (comma-separated) | `read,write` |
-
-## Next Steps
-
-- [Supabase OAuth Setup](./Supabase%20OAuth%20Setup.md) - Production auth setup
-- [External OAuth Providers](./External%20OAuth%20Providers.md) - GitHub, Google integration
-- [MCP OAuth Specification](https://modelcontextprotocol.io/specification/2025-03-26/basic/authorization) - Official spec

docs/Obsidian Integration.md

@@ -1,146 +0,0 @@
----
-title: Obsidian Integration
-type: note
-permalink: docs/obsidian-integration
----
-
-# Obsidian Integration
-
-Basic Memory integrates seamlessly with [Obsidian](https://obsidian.md), providing powerful visualization and navigation capabilities for your knowledge graph.
-
-## Setup
-
-### Creating an Obsidian Vault
-
-1. Download and install [Obsidian](https://obsidian.md)
-2. Create a new vault
-3. Point it to your Basic Memory directory (~/basic-memory by default)
-4. Enable core plugins like Graph View, Backlinks, and Tags
-
-## Visualization Features
-
-### Graph View
-
-Obsidian's Graph View provides a visual representation of your knowledge network:
-
-- Each document appears as a node
-- Relations appear as connections between nodes
-- Colors can be customized to distinguish types
-- Filters let you focus on specific aspects
-- Local graphs show connections for individual documents
-
-### Backlinks
-
-Obsidian automatically tracks references between documents:
-
-- View all documents that reference the current one
-- See the exact context of each reference
-- Navigate easily through connections
-- Track how concepts relate to each other
-
-### Tag Explorer
-
-Use tags to organize and filter content:
-
-- View all tags in your knowledge base
-- See how many documents use each tag
-- Filter documents by tag combinations
-- Create hierarchical tag structures
-
-## Knowledge Elements
-
-Basic Memory's knowledge format works natively with Obsidian:
-
-### Wiki Links
-
-```markdown
-## Relations
-- implements [[Search Design]]
-- depends_on [[Database Schema]]
-```
-
-These display as clickable links in Obsidian and appear in the graph view.
-
-### Observations with Tags
-
-```markdown
-## Observations
-- [tech] Using SQLite #database
-- [design] Local-first #architecture
-```
-
-Tags become searchable and filterable in Obsidian's tag pane.
-
-### Frontmatter
-
-```yaml
----
-title: Document Title
-type: note
-tags: [search, design]
----
-```
-
-Frontmatter provides metadata for Obsidian to use in search and filtering.
-
-## Canvas Integration
-
-Basic Memory can create [Obsidian Canvas](https://obsidian.md/canvas) files:
-
-1. Ask Claude to create a visualization:
-   ```
-   You: "Create a canvas showing the structure of our project components."
-   ```
-
-2. Claude generates a .canvas file in your knowledge base
-
-3. Open the file in Obsidian to view and edit the visual representation
-
-4. Canvas files maintain references to your documents
-
-## Recommended Plugins
-
-These Obsidian plugins work especially well with Basic Memory:
-
-- **Dataview**: Query your knowledge base programmatically
-- **Kanban**: Organize tasks from knowledge files
-- **Calendar**: View and navigate temporal knowledge
-- **Templates**: Create consistent knowledge structures
-
-## Workflow Suggestions
-
-### Daily Notes
-
-```markdown
-# 2024-01-21
-
-## Progress
-- Updated [[Search Design]]
-- Fixed [[Bug Report 123]]
-
-## Notes
-- [idea] Better indexing #enhancement
-- [todo] Update docs #documentation
-
-## Links
-- relates_to [[Current Sprint]]
-- updates [[Project Status]]
-```
-
-### Project Tracking
-
-```markdown
-# Current Sprint
-
-## Tasks
-- [ ] Update [[Search]]
-- [ ] Fix [[Auth Bug]]
-
-## Tags
-#sprint #planning #current
-```
-
-## Relations
-- enhances [[Introduction to Basic Memory]] (Overview of system)
-- relates_to [[Canvas]] (Visual knowledge mapping)
-- complements [[User Guide]] (Using Basic Memory)

docs/SPEC-PER-PROJECT-ROUTING.md

@@ -0,0 +1,175 @@
+# Per-Project Local/Cloud Routing
+
+## Context
+
+basic-memory's cloud/local mode is currently a global toggle (`cloud_mode: bool`). When enabled, ALL projects route through the cloud proxy via OAuth. This is too coarse — users should be able to keep some projects local and route others through cloud.
+
+The cloud API already supports API key auth (`bmc_`-prefixed keys, `POST /api/keys` to create, `HybridTokenVerifier` routes them automatically). API keys are per-tenant (account-level), not per-project — there are no per-project permissions in the cloud yet.
+
+**Goal**: Users can set each project to `local` or `cloud` mode. Local projects use the existing ASGI in-process transport. Cloud projects use the cloud API with a single account-level API key. No OAuth dance needed for cloud project access.
+
+## UX Flow
+
+**Option A — Create key in web app:**
+1. User creates API key in cloud web app (already supported)
+2. Copies the key
+3. Runs `bm cloud set-key bmc_abc123...` → saves to config.json
+
+**Option B — Create key via CLI:**
+1. User is already logged in via OAuth (`bm cloud login`)
+2. Runs `bm cloud create-key "my-laptop"` → calls `POST /api/keys` with JWT auth → gets key back → saves to config.json
+3. OAuth login is no longer needed for day-to-day use — the API key handles auth
+
+**Setting project mode:**
+```bash
+bm project set-cloud research      # route "research" project through cloud
+bm project set-local research      # revert to local
+bm project list                    # shows mode column (local/cloud)
+```
+
+## Implementation Plan
+
+### Step 1: Config model changes
+
+**File: `src/basic_memory/config.py`**
+
+- Add `ProjectMode` enum: `LOCAL = "local"`, `CLOUD = "cloud"`
+- Add `ProjectConfigEntry` Pydantic model: `path: str`, `mode: ProjectMode = LOCAL`
+- Evolve `BasicMemoryConfig.projects` from `Dict[str, str]` to `Dict[str, ProjectConfigEntry]`
+- Add `model_validator(mode="before")` to auto-migrate old `{"name": "/path"}` format to `{"name": {"path": "/path", "mode": "local"}}`
+- Add `cloud_api_key: Optional[str] = None` field to `BasicMemoryConfig` (account-level, not per-project)
+- Update `ProjectConfig` dataclass to carry `mode` from config entry
+- Add helpers: `get_project_entry(name)`, `get_project_mode(name)`
+- Keep global `cloud_mode` as deprecated fallback
+- Update all code that reads `config.projects` as `Dict[str, str]` to handle `ProjectConfigEntry`
+
+### Step 2: Client routing
+
+**File: `src/basic_memory/mcp/async_client.py`**
+
+- Add optional `project_name: Optional[str] = None` parameter to `get_client()`
+- Routing logic (priority order):
+  1. Factory injection (`_client_factory`) — unchanged
+  2. Force-local (`_force_local_mode()`) — unchanged
+  3. **New**: If `project_name` provided and project's mode is `CLOUD` → HTTP client with `cloud_api_key` as Bearer token, hitting `cloud_host/proxy`
+  4. Global `cloud_mode_enabled` fallback — existing OAuth flow (deprecated)
+  5. Default: local ASGI transport
+- Error if cloud project but no `cloud_api_key` in config — actionable message pointing to `bm cloud set-key` or `bm cloud create-key`
+
+### Step 3: Project-aware client helper
+
+**File: `src/basic_memory/mcp/project_context.py`**
+
+- Add `get_project_client(project, context)` async context manager
+- Combines `resolve_project_parameter()` (config-only, no network) + `get_client(project_name=resolved)` + `get_active_project(client, resolved, context)`
+- Returns `(client, active_project)` tuple
+- Solves bootstrap problem: resolve project name first, create correct client, then validate
+
+### Step 4: Simplify ProjectResolver
+
+**File: `src/basic_memory/project_resolver.py`**
+
+- Remove global `cloud_mode` parameter — routing mode is orthogonal to project resolution
+- Resolution becomes purely: constrained env var → explicit param → default project
+- Update `resolve_project_parameter()` in `project_context.py` to drop `cloud_mode` param
+
+### Step 5: Update MCP tools
+
+**Files: `src/basic_memory/mcp/tools/*.py` (~15 files)**
+
+Mechanical change per tool:
+```python
+# Before
+async with get_client() as client:
+    active_project = await get_active_project(client, project, context)
+
+# After
+async with get_project_client(project, context) as (client, active_project):
+```
+
+Special handling for `recent_activity.py` discovery mode: iterate projects, create per-project client for each.
+
+### Step 6: Sync coordinator
+
+**Files: `src/basic_memory/sync/coordinator.py`, `src/basic_memory/mcp/container.py`**
+
+- Filter file watchers to local-mode projects only
+- Cloud projects skip sync
+
+### Step 7: CLI commands
+
+**File: `src/basic_memory/cli/commands/cloud/core_commands.py`**
+
+- `bm cloud set-key <api-key>` — saves API key to config.json
+- `bm cloud create-key <name>` — calls `POST {cloud_host}/api/keys` using existing JWT auth (from `make_api_request`), saves returned key to config. Uses existing `api_client.py:make_api_request()` for the authenticated call.
+
+**File: `src/basic_memory/cli/commands/project.py`**
+
+- `bm project set-cloud <name>` — sets project mode to cloud (validates API key exists in config)
+- `bm project set-local <name>` — reverts project to local mode
+- Extend `bm project list` / `bm project info` to show mode column
+
+### Step 8: RuntimeMode simplification
+
+**File: `src/basic_memory/runtime.py`**
+
+- `resolve_runtime_mode()` drops `cloud_mode_enabled` parameter
+- Simplifies to: TEST if test env, otherwise LOCAL
+- `RuntimeMode.CLOUD` kept for backward compat but not used in global resolution
+
+### Step 9: Tests
+
+- Config: migration from old format, round-trip serialization, `get_project_mode()`
+- `get_client()`: local project → ASGI, cloud project → HTTP+API key, missing key → error
+- `get_project_client()`: resolve + route combined
+- MCP tools: representative sample with new helper
+- Sync: cloud projects skipped, local projects synced
+- CLI: `set-key`, `create-key`, `set-cloud`, `set-local`
+
+## Key Files
+
+| File | Change |
+|------|--------|
+| `src/basic_memory/config.py` | `ProjectMode`, `ProjectConfigEntry`, migration, `cloud_api_key` field |
+| `src/basic_memory/mcp/async_client.py` | `get_client(project_name=)` per-project routing |
+| `src/basic_memory/mcp/project_context.py` | `get_project_client()` helper |
+| `src/basic_memory/project_resolver.py` | Remove global `cloud_mode` concern |
+| `src/basic_memory/mcp/tools/*.py` | Mechanical swap to `get_project_client()` |
+| `src/basic_memory/sync/coordinator.py` | Filter to local-mode projects |
+| `src/basic_memory/mcp/container.py` | Update should_sync logic |
+| `src/basic_memory/cli/commands/cloud/core_commands.py` | `set-key`, `create-key` commands |
+| `src/basic_memory/cli/commands/project.py` | `set-cloud`, `set-local` commands |
+| `src/basic_memory/runtime.py` | Drop cloud_mode from global resolution |
+
+## Config Example
+
+```json
+{
+  "projects": {
+    "personal": {"path": "/Users/me/notes", "mode": "local"},
+    "research": {"path": "/Users/me/research", "mode": "cloud"}
+  },
+  "cloud_api_key": "bmc_abc123...",
+  "cloud_host": "https://cloud.basicmemory.com",
+  "default_project": "personal"
+}
+```
+
+## Edge Cases
+
+| Case | Handling |
+|------|----------|
+| No API key + cloud project | `get_client()` raises error: "Run `bm cloud set-key` first" |
+| Old config format loaded | `model_validator` auto-migrates `Dict[str,str]` to new format |
+| Default project is cloud | Works — resolver returns name, routing uses API key |
+| Global `cloud_mode=true` (legacy) | Deprecated fallback still works via OAuth |
+| Factory-injected client (cloud app) | Factory takes priority, unaffected |
+| `--local` CLI flag on cloud project | Force-local override still works |
+
+## Verification
+
+1. `just fast-check` — lint/format/typecheck + impacted tests
+2. `just test` — full suite (SQLite + Postgres)
+3. Manual: `bm cloud set-key bmc_...`, `bm project set-cloud test`, run MCP tools against it
+4. Manual: verify local projects work unchanged
+5. Manual: `bm project list` shows mode column

docs/Supabase OAuth Setup.md

@@ -1,311 +0,0 @@
-# Supabase OAuth Setup for Basic Memory
-
-This guide explains how to set up Supabase as the OAuth provider for Basic Memory MCP server in production.
-
-## Prerequisites
-
-1. A Supabase project (create one at [supabase.com](https://supabase.com))
-2. Basic Memory MCP server deployed
-3. Environment variables configuration
-
-## Overview
-
-The Supabase OAuth provider offers:
-- Production-ready authentication with persistent storage
-- User management through Supabase Auth
-- JWT token validation
-- Integration with Supabase's security features
-- Support for social logins (GitHub, Google, etc.)
-
-## Setup Steps
-
-### 1. Get Supabase Credentials
-
-From your Supabase project dashboard:
-
-1. Go to Settings > API
-2. Copy these values:
-   - `Project URL` → `SUPABASE_URL`
-   - `anon public` key → `SUPABASE_ANON_KEY`
-   - `service_role` key → `SUPABASE_SERVICE_KEY` (keep this secret!)
-   - JWT secret → `SUPABASE_JWT_SECRET` (under Settings > API > JWT Settings)
-
-### 2. Configure Environment Variables
-
-Create a `.env` file:
-
-```bash
-# Enable OAuth
-FASTMCP_AUTH_ENABLED=true
-FASTMCP_AUTH_PROVIDER=supabase
-
-# Your MCP server URL
-FASTMCP_AUTH_ISSUER_URL=https://your-mcp-server.com
-
-# Supabase configuration
-SUPABASE_URL=https://your-project.supabase.co
-SUPABASE_ANON_KEY=your-anon-key
-SUPABASE_SERVICE_KEY=your-service-key
-SUPABASE_JWT_SECRET=your-jwt-secret
-
-# Allowed OAuth clients (comma-separated)
-SUPABASE_ALLOWED_CLIENTS=web-app,mobile-app,cli-tool
-
-# Required scopes
-FASTMCP_AUTH_REQUIRED_SCOPES=read,write
-```
-
-### 3. Create OAuth Clients Table (Optional)
-
-For production, create a table to store OAuth clients in Supabase:
-
-```sql
-CREATE TABLE oauth_clients (
-  id UUID DEFAULT gen_random_uuid() PRIMARY KEY,
-  client_id TEXT UNIQUE NOT NULL,
-  client_secret TEXT NOT NULL,
-  name TEXT,
-  redirect_uris TEXT[],
-  allowed_scopes TEXT[],
-  created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW(),
-  updated_at TIMESTAMP WITH TIME ZONE DEFAULT NOW()
-);
-
--- Create an index for faster lookups
-CREATE INDEX idx_oauth_clients_client_id ON oauth_clients(client_id);
-
--- RLS policies
-ALTER TABLE oauth_clients ENABLE ROW LEVEL SECURITY;
-
--- Only service role can manage clients
-CREATE POLICY "Service role can manage clients" ON oauth_clients
-  FOR ALL USING (auth.jwt()->>'role' = 'service_role');
-```
-
-### 4. Set Up Auth Flow
-
-The Supabase OAuth provider handles the following flow:
-
-1. **Client Authorization Request**
-   ```
-   GET /authorize?client_id=web-app&redirect_uri=https://app.com/callback
-   ```
-
-2. **Redirect to Supabase Auth**
-   - User authenticates with Supabase (email/password, magic link, or social login)
-   - Supabase redirects back to your MCP server
-
-3. **Token Exchange**
-   ```
-   POST /token
-   Content-Type: application/x-www-form-urlencoded
-   
-   grant_type=authorization_code&code=xxx&client_id=web-app
-   ```
-
-4. **Access Protected Resources**
-   ```
-   GET /mcp
-   Authorization: Bearer <access_token>
-   ```
-
-### 5. Enable Social Logins (Optional)
-
-In Supabase dashboard:
-
-1. Go to Authentication > Providers
-2. Enable desired providers (GitHub, Google, etc.)
-3. Configure OAuth apps for each provider
-4. Users can now log in via social providers
-
-### 6. User Management
-
-Supabase provides:
-- User registration and login
-- Password reset flows
-- Email verification
-- User metadata storage
-- Admin APIs for user management
-
-Access user data in your MCP tools:
-
-```python
-# In your MCP tool
-async def get_user_info(ctx: Context):
-    # The token is already validated by the OAuth middleware
-    user_id = ctx.auth.user_id
-    email = ctx.auth.email
-    
-    # Use Supabase client to get more user data if needed
-    user = await supabase.auth.admin.get_user_by_id(user_id)
-    return user
-```
-
-### 7. Production Deployment
-
-1. **Environment Security**
-   - Never expose `SUPABASE_SERVICE_KEY`
-   - Use environment variables, not hardcoded values
-   - Rotate keys periodically
-
-2. **HTTPS Required**
-   - Always use HTTPS in production
-   - Configure proper SSL certificates
-
-3. **Rate Limiting**
-   - Implement rate limiting for auth endpoints
-   - Use Supabase's built-in rate limiting
-
-4. **Monitoring**
-   - Monitor auth logs in Supabase dashboard
-   - Set up alerts for suspicious activity
-
-## Testing
-
-### Local Development
-
-For local testing with Supabase:
-
-```bash
-# Start MCP server with Supabase auth
-FASTMCP_AUTH_ENABLED=true \
-FASTMCP_AUTH_PROVIDER=supabase \
-SUPABASE_URL=http://localhost:54321 \
-SUPABASE_ANON_KEY=your-local-anon-key \
-bm mcp --transport streamable-http
-```
-
-### Test Authentication Flow
-
-```python
-import httpx
-import asyncio
-
-async def test_supabase_auth():
-    # 1. Register/login with Supabase directly
-    supabase_url = "https://your-project.supabase.co"
-    
-    # 2. Get MCP authorization URL
-    response = await httpx.get(
-        "http://localhost:8000/authorize",
-        params={
-            "client_id": "web-app",
-            "redirect_uri": "http://localhost:3000/callback",
-            "response_type": "code",
-        }
-    )
-    
-    # 3. User logs in via Supabase
-    # 4. Exchange code for MCP tokens
-    # 5. Access protected resources
-
-asyncio.run(test_supabase_auth())
-```
-
-## Advanced Configuration
-
-### Custom User Metadata
-
-Store additional user data in Supabase:
-
-```sql
--- Add custom fields to auth.users
-ALTER TABLE auth.users 
-ADD COLUMN IF NOT EXISTS metadata JSONB DEFAULT '{}';
-
--- Or create a separate profiles table
-CREATE TABLE profiles (
-  id UUID REFERENCES auth.users PRIMARY KEY,
-  username TEXT UNIQUE,
-  avatar_url TEXT,
-  bio TEXT,
-  created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW()
-);
-```
-
-### Row Level Security (RLS)
-
-Protect user data with RLS:
-
-```sql
--- Users can only access their own data
-CREATE POLICY "Users can view own profile" ON profiles
-  FOR SELECT USING (auth.uid() = id);
-
-CREATE POLICY "Users can update own profile" ON profiles
-  FOR UPDATE USING (auth.uid() = id);
-```
-
-### Custom Claims
-
-Add custom claims to JWT tokens:
-
-```sql
--- Function to add custom claims
-CREATE OR REPLACE FUNCTION custom_jwt_claims()
-RETURNS JSON AS $$
-BEGIN
-  RETURN json_build_object(
-    'user_role', current_setting('request.jwt.claims')::json->>'user_role',
-    'permissions', current_setting('request.jwt.claims')::json->>'permissions'
-  );
-END;
-$$ LANGUAGE plpgsql;
-```
-
-## Troubleshooting
-
-### Common Issues
-
-1. **Invalid JWT Secret**
-   - Ensure `SUPABASE_JWT_SECRET` matches your Supabase project
-   - Check Settings > API > JWT Settings in Supabase dashboard
-
-2. **CORS Errors**
-   - Configure CORS in your MCP server
-   - Add allowed origins in Supabase dashboard
-
-3. **Token Validation Fails**
-   - Verify tokens are being passed correctly
-   - Check token expiration times
-   - Ensure scopes match requirements
-
-4. **User Not Found**
-   - Confirm user exists in Supabase Auth
-   - Check if email is verified (if required)
-   - Verify client permissions
-
-### Debug Mode
-
-Enable debug logging:
-
-```bash
-export FASTMCP_LOG_LEVEL=DEBUG
-export SUPABASE_LOG_LEVEL=debug
-```
-
-## Security Best Practices
-
-1. **Secure Keys**: Never commit secrets to version control
-2. **Least Privilege**: Use minimal required scopes
-3. **Token Rotation**: Implement refresh token rotation
-4. **Audit Logs**: Monitor authentication events
-5. **Rate Limiting**: Protect against brute force attacks
-6. **HTTPS Only**: Always use encrypted connections
-
-## Migration from Basic Auth
-
-To migrate from the basic auth provider:
-
-1. Export existing user data
-2. Import users into Supabase Auth
-3. Update client applications to use new auth flow
-4. Gradually transition users to Supabase login
-
-## Next Steps
-
-- Set up email templates in Supabase
-- Configure password policies
-- Implement MFA (multi-factor authentication)
-- Add social login providers
-- Create admin dashboard for user management

docs/Technical Information.md

@@ -1,243 +0,0 @@
----
-title: Technical Information
-type: note
-permalink: docs/technical-information
----
-
-# Technical Information
-
-This document provides technical details about Basic Memory's implementation, licensing, and integration with the Model Context Protocol (MCP).
-
-## Architecture
-
-Basic Memory consists of:
-
-1. **Core Knowledge Engine**: Parses and indexes Markdown files
-2. **SQLite Database**: Provides fast querying and search
-3. **MCP Server**: Implements the Model Context Protocol
-4. **CLI Tools**: Command-line utilities for management
-5. **Sync Service**: Monitors file changes and updates the database
-
-The system follows a file-first architecture where all knowledge is represented in standard Markdown files and the database serves as a secondary index.
-
-## Model Context Protocol (MCP)
-
-Basic Memory implements the [Model Context Protocol](https://github.com/modelcontextprotocol/spec), an open standard for enabling AI models to access external tools:
-
-- **Standardized Interface**: Common protocol for tool integration
-- **Tool Registration**: Basic Memory registers as a tool provider
-- **Asynchronous Communication**: Enables efficient interaction with AI models
-- **Standardized Schema**: Structured data exchange format
-
-Integration with Claude Desktop uses the MCP to grant Claude access to your knowledge base through a set of specialized tools that search, read, and write knowledge.
-
-## Licensing
-
-Basic Memory is licensed under the [GNU Affero General Public License v3.0 (AGPL-3.0)](https://www.gnu.org/licenses/agpl-3.0.en.html):
-
-- **Free Software**: You can use, study, share, and modify the software
-- **Copyleft**: Derivative works must be distributed under the same license
-- **Network Use**: Network users must be able to receive the source code
-- **Commercial Use**: Allowed, subject to license requirements
-
-The AGPL license ensures Basic Memory remains open source while protecting against proprietary forks.
-
-## Source Code
-
-Basic Memory is developed as an open-source project:
-
-- **GitHub Repository**: [https://github.com/basicmachines-co/basic-memory](https://github.com/basicmachines-co/basic-memory)
-- **Issue Tracker**: Report bugs and request features on GitHub
-- **Contributions**: Pull requests are welcome following the contributing guidelines
-- **Documentation**: Source for this documentation is also available in the repository
-
-## Data Storage and Privacy
-
-Basic Memory is designed with privacy as a core principle:
-
-- **Local-First**: All data remains on your local machine
-- **No Cloud Dependency**: No remote servers or accounts required
-- **Telemetry**: Optional and disabled by default
-- **Standard Formats**: All data is stored in standard file formats you control
-
-## Implementation Details
-
-Knowledge in Basic Memory is organized as a semantic graph:
-
-1. **Entities** - Distinct concepts represented by Markdown documents
-2. **Observations** - Categorized facts and information about entities
-3. **Relations** - Connections between entities that form the knowledge graph
-
-This structure emerges from simple text patterns in standard Markdown:
-
-```markdown
----
-title: Coffee Brewing Methods
-type: note
-permalink: coffee/coffee-brewing-methods
-tags:
-- '#coffee'
-- '#brewing'
-- '#methods'
-- '#demo'
----
-
-# Coffee Brewing Methods
-
-An exploration of different coffee brewing techniques, their characteristics, and how they affect flavor extraction.
-
-## Overview
-
-Coffee brewing is both an art and a science. Different brewing methods extract different compounds from coffee beans,
-resulting in unique flavor profiles, body, and mouthfeel. The key variables in any brewing method are:
-
-- Grind size
-- Water temperature
-- Brew time
-- Coffee-to-water ratio
-- Agitation/turbulence
-
-## Observations
-
-- [principle] Coffee extraction follows a predictable pattern: acids extract first, then sugars, then bitter compounds
-  #extraction
-- [method] Pour over methods generally produce cleaner, brighter cups with more distinct flavor notes #clarity
-
-## Relations
-
-- requires [[Proper Grinding Technique]]
-- affects [[Flavor Extraction]]
-```
-
-Becomes
-
-```json
-{
-  "entities": [
-    {
-      "permalink": "coffee/coffee-brewing-methods",
-      "title": "Coffee Brewing Methods",
-      "file_path": "Coffee Notes/Coffee Brewing Methods.md",
-      "entity_type": "note",
-      "entity_metadata": {
-        "title": "Coffee Brewing Methods",
-        "type": "note",
-        "permalink": "coffee/coffee-brewing-methods",
-        "tags": "['#coffee', '#brewing', '#methods', '#demo']"
-      },
-      "checksum": "bfa32a0f23fa124b53f0694c344d2788b0ce50bd090b55b6d738401d2a349e4c",
-      "content_type": "text/markdown",
-      "observations": [
-        {
-          "category": "principle",
-          "content": "Coffee extraction follows a predictable pattern: acids extract first, then sugars, then bitter compounds #extraction",
-          "tags": [
-            "extraction"
-          ],
-          "permalink": "coffee/coffee-brewing-methods/observations/principle/coffee-extraction-follows-a-predictable-pattern-acids-extract-first-then-sugars-then-bitter-compounds-extraction"
-        },
-        {
-          "category": "method",
-          "content": "Pour over methods generally produce cleaner, brighter cups with more distinct flavor notes #clarity",
-          "tags": [
-            "clarity"
-          ],
-          "permalink": "coffee/coffee-brewing-methods/observations/method/pour-over-methods-generally-produce-cleaner-brighter-cups-with-more-distinct-flavor-notes-clarity"
-        }
-      ],
-      "relations": [
-        {
-          "from_id": "coffee/coffee-bean-origins",
-          "to_id": "coffee/coffee-brewing-methods",
-          "relation_type": "pairs_with",
-          "permalink": "coffee/coffee-bean-origins/pairs-with/coffee/coffee-brewing-methods",
-          "to_name": "Coffee Brewing Methods"
-        },
-        {
-          "from_id": "coffee/flavor-extraction",
-          "to_id": "coffee/coffee-brewing-methods",
-          "relation_type": "affected_by",
-          "permalink": "coffee/flavor-extraction/affected-by/coffee/coffee-brewing-methods",
-          "to_name": "Coffee Brewing Methods"
-        }
-      ],
-      "created_at": "2025-03-06T14:01:23.445071",
-      "updated_at": "2025-03-06T13:34:48.563606"
-    }
-  ]
-}
-```
-
-Basic Memory understands how to build context via its semantic graph.
-
-### Entity Model
-
-Basic Memory's core data model consists of:
-
-- **Entities**: Documents in your knowledge base
-- **Observations**: Facts or statements about entities
-- **Relations**: Connections between entities
-- **Tags**: Additional categorization for entities and observations
-
-The system parses Markdown files to extract this structured information while preserving the human-readable format.
-
-### Files as Source of Truth
-
-Plain Markdown files store all knowledge, making it accessible with any text editor and easy to version with git.
-
-```mermaid
-flowchart TD
-    User((User)) <--> |Conversation| Claude["Claude or other LLM"]
-    Claude <-->|API Calls| BMCP["Basic Memory MCP Server"]
-    
-    subgraph "Local Storage"
-        KnowledgeFiles["Markdown Files - Source of Truth"]
-        KnowledgeIndex[(Knowledge Graph SQLite Index)]
-    end
-    
-    BMCP <-->|"write_note() read_note()"| KnowledgeFiles
-    BMCP <-->|"search_notes() build_context()"| KnowledgeIndex
-    KnowledgeFiles <-.->|Sync Process| KnowledgeIndex
-    KnowledgeFiles <-->|Direct Editing| Editors((Text Editors & Git))
-    
-    User -.->|"Complete control, Privacy preserved"| KnowledgeFiles
-    
-    
-    class Claude primary
-    class BMCP secondary
-    class KnowledgeFiles tertiary
-    class KnowledgeIndex quaternary
-    class User,Editors user`;
-```
-
-### Sqlite Database
-
-A local SQLite database maintains the knowledge graph topology for fast queries and semantic traversal without cloud dependencies. It contains:
-- db tables for the knowledge graph schema
-- a search index table enabling full text search across the knowledge base
-
-
-### Sync Process
-
-The sync process:
-
-1. Detects changes to files in the knowledge directory
-2. Parses modified files to extract structured data
-3. Updates the SQLite database with changes
-4. Resolves forward references when new entities are created
-5. Updates the search index for fast querying
-
-### Search Engine
-
-The search functionality:
-
-1. Uses a combination of full-text search and semantic matching
-2. Indexes observations, relations, and content
-3. Supports wildcards and pattern matching in memory:// URLs
-4. Traverses the knowledge graph to follow relationships
-5. Ranks results by relevance to the query
-
-## Relations
-- relates_to [[Welcome to Basic memory]] (Overview)
-- relates_to [[CLI Reference]] (Command line tools)
-- implements [[Knowledge Format]] (File structure and format)

docs/User Guide.md

@@ -1,657 +0,0 @@
----
-title: User Guide
-type: note
-permalink: docs/user-guide
----
-
-# User Guide
-
-This guide explains how to effectively use Basic Memory in your daily workflow, from creating knowledge through
-conversations to building a rich semantic network.
-
-## Basic Memory Workflow
-
-Using Basic Memory follows a natural cycle:
-
-1. **Have conversations** with AI assistants like Claude
-2. **Capture knowledge** in Markdown files
-3. **Build connections** between pieces of knowledge
-4. **Reference your knowledge** in future conversations
-5. **Edit files directly** when needed
-6. **Sync changes** automatically
-
-## Creating Knowledge
-
-### Through Conversations
-
-To create knowledge during conversations with Claude:
-
-```
-You: We've covered several authentication approaches. Could you create a note summarizing what we've discussed?
-
-Claude: I'll create a note summarizing our authentication discussion.
-```
-
-This creates a Markdown file in your `~/basic-memory` directory with semantic markup.
-
-### Direct File Creation
-
-You can create files directly:
-
-1. Create a new Markdown file in your `~/basic-memory` directory
-2. Add frontmatter with title, type, and optional tags
-3. Structure content with observations and relations
-4. Save the file
-5. Run `basic-memory sync` if not in watch mode
-
-## Using Special Prompts
-
-Basic Memory includes several special prompts that help you leverage your knowledge base more effectively. In apps like
-Claude Desktop, these prompts trigger specific tools to search and analyze your knowledge base.
-
-### Continue Conversation
-
-When you want to pick up where you left off on a topic:
-
-```
-You: Let's continue our conversation about authentication systems.
-```
-
-Behind the scenes:
-
-- Claude searches your knowledge base for content about "authentication systems"
-- It retrieves relevant documents and their relations
-- It analyzes the context to understand where you left off
-- It builds a comprehensive picture of what you've previously discussed
-- It can then resume the conversation with all that context
-
-This is particularly useful when:
-
-- Starting a new session days or weeks after your last discussion
-- Switching between multiple ongoing projects
-- Building on previous work without repeating yourself
-
-### Recent Activity
-
-To get an overview of what you've been working on:
-
-```
-You: What have we been discussing recently?
-```
-
-Behind the scenes:
-
-- Claude retrieves documents modified recently
-- It analyzes patterns and themes
-- It summarizes the key topics and changes
-- It offers to continue working on any of those topics
-
-This is useful for:
-
-- Coming back after a break
-- Getting a quick reminder of ongoing projects
-- Deciding what to work on next
-
-### Search
-
-To find specific information in your knowledge base:
-
-```
-You: Find information about JWT authentication in my notes.
-```
-
-Behind the scenes:
-
-- Claude performs a semantic search for "JWT authentication"
-- It retrieves and ranks the most relevant documents
-- It summarizes the key findings
-- It offers to explore specific areas in more detail
-
-This is useful for:
-
-- Finding specific information quickly
-- Exploring what you know about a topic
-- Starting work on an existing topic
-
-### Example
-
-Choose "Continue Conversation"
-![[prompt 1.png|500]]
-
-Enter a topic
-![[prompt2.png|500]]
-
-Give instructions
-![[prompt3.png|500]]
-
-Claude Desktop lets you send a prompt to provide context. You can use this at the beginning of a chat to preload context
-without needing to copy paste all the time. By using one of the supplied prompts, Basic Memory will search the knowledge
-base and give the AI instructions for how to build context.
-
-Choose "Continue Conversation":
-
-![[prompt 1.png|500]]
-
-Enter a topic:
-
-![[prompt2.png|500]]
-
-Give optional additional instructions:
-
-![[prompt3.png|500]]
-
-Claude can build context from the supplied topic. This works independently of Claude Project information. All the
-context comes from your local knowledge base.
-
-![[prompt4.png|500]]
-
-## Searching Your Knowledge Base
-
-Basic Memory provides multiple ways to search and explore your knowledge base:
-
-### Natural Language Search
-
-The simplest way to search is to ask Claude directly:
-
-```
-You: What do I know about authentication methods?
-```
-
-Claude will search your knowledge base semantically and return relevant information.
-
-### Search Prompt
-
-Use the dedicated search prompt for more focused searches:
-
-```
-You: Search for "JWT authentication"
-```
-
-This triggers a specialized search that returns precise results with document titles, relevant excerpts, and offers to
-explore specific documents.
-
-### Boolean Search
-
-For more precise searches, use boolean operators to refine your queries:
-
-```
-You: Search for "authentication AND OAuth NOT basic"
-```
-
-Basic Memory supports standard boolean operators:
-
-- **AND**: Find documents containing both terms
-  ```
-  You: Search for "python AND flask"
-  ```
-  This finds documents containing both "python" and "flask"
-
-- **OR**: Find documents containing either term
-  ```
-  You: Search for "python OR javascript"
-  ```
-  This finds documents containing either "python" or "javascript"
-
-- **NOT**: Exclude documents containing specific terms
-  ```
-  You: Search for "python NOT django"
-  ```
-  This finds documents containing "python" but excludes those containing "django"
-
-- **Grouping with parentheses**: Control operator precedence
-  ```
-  You: Search for "(python OR javascript) AND web"
-  ```
-  This finds documents about web development that mention either Python or JavaScript
-
-Boolean search is particularly useful for:
-
-- Narrowing down results in large knowledge bases
-- Finding specific combinations of concepts
-- Excluding irrelevant content from search results
-- Creating complex queries for precise information retrieval
-
-### Memory URL Pattern Matching
-
-For advanced searches, use memory:// URL patterns with wildcards:
-
-```
-You: Look at memory://auth* and summarize all authentication approaches.
-```
-
-Pattern matching supports:
-
-- **Wildcards**: `memory://auth*` matches all permalinks starting with "auth"
-- **Path patterns**: `memory://project/*/auth` matches auth documents in any project subfolder
-- **Relation traversal**: `memory://auth-system/implements/*` finds all documents that implement the auth system
-
-### Combining Search with Context Building
-
-The most powerful searches build comprehensive context by following relationships:
-
-```
-You: Search for JWT authentication and then follow all implementation relations.
-```
-
-This builds a complete picture by:
-
-1. Finding documents about JWT authentication
-2. Following implementation relationships from those documents
-3. Building a complete picture of how JWT is implemented across your system
-
-### Search Best Practices
-
-For effective searching:
-
-1. **Be specific** with search terms and phrases
-2. **Use boolean operators** to refine searches and find precise information
-3. **Use technical terms** when searching for technical content
-4. **Follow up** on search results by asking for more details about specific documents
-5. **Combine approaches** by starting with search and then using memory:// URLs for precision
-6. **Use relation traversal** to explore connected concepts after finding initial documents
-
-## Referencing Knowledge
-
-### Using memory:// URLs
-
-Reference specific knowledge directly:
-
-```
-You: Please look at memory://authentication-approaches and suggest which approach would be best for our mobile app.
-```
-
-### Natural Language References
-
-Reference knowledge conversationally:
-
-```
-You: What did we decide about authentication for the project?
-```
-
-### Advanced References
-
-Follow connections across your knowledge graph:
-
-```
-You: Look at memory://project-architecture and check related documents to give me a complete picture.
-```
-
-## Working with Files
-
-### File Location and Organization
-
-By default, Basic Memory stores files in `~/basic-memory`:
-
-- Browse this directory in your file explorer
-- Organize files into subfolders
-- Use git for version control
-
-### File Format
-
-Each knowledge file follows this structure:
-
-```markdown
----
-title: Authentication Approaches
-type: note
-tags: [security, architecture]
-permalink: authentication-approaches
----
-
-# Authentication Approaches
-
-A comparison of authentication methods.
-
-## Observations
-
-- [approach] JWT provides stateless authentication #security
-- [limitation] Session tokens require server-side storage #infrastructure
-
-## Relations
-
-- implements [[Security Requirements]]
-- affects [[User Login Flow]]
-```
-
-### Editing Files
-
-Modify files in any text editor:
-
-1. Open the file in your preferred editor
-2. Make changes to content, observations, or relations
-3. Save the file
-4. Basic Memory detects changes automatically when running in watch mode
-
-## Building a Knowledge Graph
-
-The value of Basic Memory comes from connections between pieces of knowledge.
-
-### Creating Relations
-
-When creating or editing notes, build connections:
-
-```markdown
-## Relations
-
-- implements [[Security Requirements]]
-- depends_on [[User Authentication]]
-```
-
-Relations can be:
-
-- Hierarchical (part_of, contains)
-- Directional (implements, depends_on)
-- Associative (relates_to, similar_to)
-- Temporal (precedes, follows)
-
-Relations are also created via regular wiki-link style links within the body text.
-
-### Forward References
-
-Reference documents that don't exist yet:
-
-```markdown
-- will_impact [[Future Feature]]
-```
-
-These references resolve automatically when you create the referenced document.
-
-## Conversation Continuity
-
-Basic Memory maintains context across different conversations.
-
-### Starting New Sessions with Context
-
-When starting a new conversation with Claude, you can:
-
-1. **Use special prompts** like "Continue conversation about..." or "What were we working on?"
-2. **Reference specific documents** with memory:// URLs
-3. **Ask about recent work** with "What have we been discussing recently?"
-4. **Search for specific topics** with "Find information about..."
-
-### Long-Term Projects
-
-Maintain context for complex projects over time:
-
-1. **Document key decisions** as you make them
-2. **Create relationships** between project components
-3. **Reference past decisions** when implementing features
-4. **Update documentation** as the project evolves
-
-### Tips for Effective Continuity
-
-1. **Be specific about topics** when continuing a conversation
-2. **Reference documents directly** with memory:// URLs for precision
-3. **Create summary notes** after important discussions
-4. **Update existing notes** rather than creating duplicates
-5. **Build robust connections** between related topics
-
-## Advanced Features
-
-### Note Editing (New in v0.13.0)
-
-**Edit notes incrementally without rewriting entire documents:**
-
-```
-💬 "Add a new section about deployment to my API documentation"
-🤖 [Uses edit_note to append new section]
-
-💬 "Update the date at the top of my meeting notes"
-🤖 [Uses edit_note to prepend new timestamp]
-
-💬 "Replace the implementation section in my design doc"
-🤖 [Uses edit_note to replace specific section]
-```
-
-Available editing operations:
-- **Append**: Add content to end of notes
-- **Prepend**: Add content to beginning of notes
-- **Replace Section**: Replace content under specific headers
-- **Find & Replace**: Simple text replacements with validation
-
-### File Management (New in v0.13.0)
-
-**Move and organize notes with full database consistency:**
-
-```
-💬 "Move my old meeting notes to the archive folder"
-🤖 [Uses move_note with automatic folder creation and database updates]
-
-💬 "Reorganize my project files into a better structure"
-🤖 [Moves files while maintaining search indexes and links]
-```
-
-Move operations include:
-- **Database Consistency**: Updates file paths, permalinks, and checksums
-- **Search Reindexing**: Maintains search functionality after moves
-- **Folder Creation**: Automatically creates destination directories
-- **Project Isolation**: Moves are contained within the current project
-- **Rollback Protection**: Ensures data integrity during failed operations
-
-### Enhanced Search (New in v0.13.0)
-
-**Frontmatter tags are now searchable:**
-
-```yaml
----
-title: Coffee Brewing Methods
-tags: [coffee, brewing, equipment]
----
-```
-
-Now searchable by: "coffee", "brewing", "equipment", or "Coffee Brewing Methods"
-
-### Importing External Knowledge
-
-Import existing conversations:
-
-```bash
-# From Claude
-basic-memory import claude conversations
-
-# From ChatGPT
-basic-memory import chatgpt
-
-# Target specific projects (v0.13.0)
-basic-memory --project=work import claude conversations
-```
-
-After importing, changes sync automatically in real-time.
-
-### Obsidian Integration
-
-Use with [Obsidian](https://obsidian.md):
-
-1. Point Obsidian to your `~/basic-memory` directory
-2. Use Obsidian's graph view to visualize your knowledge network
-3. All changes sync back to Basic Memory
-
-### Canvas Visualizations
-
-Create visual knowledge maps:
-
-```
-You: Could you create a canvas visualization of our project components?
-```
-
-This generates an Obsidian canvas file showing the relationships between concepts.
-
-### Advanced Memory URI Patterns
-
-Use wildcards and patterns:
-
-```
-You: Review memory://project/*/requirements to summarize all project requirements.
-```
-
-## Command Line Interface
-
-### Sync Commands
-
-```bash
-# One-time sync
-basic-memory sync
-
-# Watch for changes
-basic-memory sync --watch
-```
-
-### Status and Information
-
-```bash
-# Check system status
-basic-memory status
-
-# View CLI help
-basic-memory --help
-```
-
-### Import Commands
-
-```bash
-# Import from Claude
-basic-memory import claude conversations
-
-# Import from ChatGPT
-basic-memory import chatgpt
-```
-
-## Multiple Projects (v0.13.0)
-
-Basic Memory v0.13.0 introduces **fluid project management** - the ability to switch between projects instantly during conversations without restart. This allows you to maintain separate knowledge graphs for different purposes while seamlessly switching between them.
-
-### Instant Project Switching (New in v0.13.0)
-
-**Switch projects during conversations:**
-
-```
-💬 "What projects do I have?"
-🤖 Available projects:
-   • main (current, default)
-   • work-notes
-   • personal-journal
-   • code-snippets
-
-💬 "Switch to work-notes"
-🤖 ✓ Switched to work-notes project
-   
-   Project Summary:
-   • 47 entities
-   • 125 observations  
-   • 23 relations
-
-💬 "What did I work on yesterday?"
-🤖 [Shows recent activity from work-notes project]
-```
-
-### Project-Specific Operations (New in v0.13.0)
-
-Some MCP tools support optional project parameters for targeting specific projects:
-
-```
-💬 "Create a note about this meeting in my personal-notes project"
-🤖 [Creates note in personal-notes project]
-
-💬 "Switch to my work project"
-🤖 [Switches project context, then all operations work within that project]
-```
-
-**Note**: Operations like search, move, and edit work within the currently active project. To work with content in different projects, switch to that project first or use the project parameter where supported.
-
-### Managing Projects
-
-```bash  
-# List all configured projects  
-basic-memory project list  
-  
-# Add a new project  
-basic-memory project create work ~/work-basic-memory  
-  
-# Set the default project  
-basic-memory project set-default work  
-  
-# Remove a project (doesn't delete files)  
-basic-memory project delete personal  
-  
-# Show current project statistics
-basic-memory project info
-```  
-
-### Using Projects in Commands
-
-All commands support the `--project` flag to specify which project to use:
-
-```bash  
-# Sync a specific project  
-basic-memory --project=work sync  
-  
-# Run MCP server for a specific project  
-basic-memory --project=personal mcp  
-```  
-
-You can also set the `BASIC_MEMORY_PROJECT` environment variable:
-
-```bash  
-BASIC_MEMORY_PROJECT=work basic-memory sync  
-```  
-
-### Unified Database Architecture (New in v0.13.0)
-
-Basic Memory v0.13.0 uses a unified database architecture:
-
-- **Single Database**: All projects share `~/.basic-memory/memory.db`
-- **Project Isolation**: Proper data separation with project context
-- **Better Performance**: Optimized queries and reduced file I/O
-- **Easier Backup**: Single database file contains all project data
-- **Session Context**: Maintains active project throughout conversations
-
-## Workflow Tips
-
-### General Workflow
-1. **Project Organization**: Use multiple projects to separate different areas (work, personal, research)
-2. **Session Context**: Switch projects during conversations without restart (v0.13.0)
-3. **Real-time Sync**: Changes sync automatically - no need to run watch mode
-4. **Review Content**: Edit AI-created content for accuracy
-5. **Build Connections**: Create rich relationships between related ideas
-6. **Use Special Prompts**: Start conversations with context from your knowledge base
-
-### v0.13.0 Workflow Enhancements
-7. **Incremental Editing**: Use edit_note for small changes instead of rewriting entire documents
-8. **File Organization**: Move and reorganize notes as your knowledge base grows
-9. **Project-Specific Creation**: Create notes in specific projects using project parameters
-10. **Search Tags**: Use frontmatter tags to improve content discoverability
-11. **Project Statistics**: Monitor project growth and activity with project info commands
-
-## Troubleshooting
-
-### Sync Issues
-
-If changes aren't showing up:
-
-1. Run `basic-memory status` to check system state 
-2. Try a manual sync with `basic-memory sync`
-
-### Missing Content
-
-If content isn't found:
-
-1. Check the exact path and permalink
-2. Try searching with more general terms
-3. Verify the file exists in your knowledge base
-
-### Relation Problems
-
-If relations aren't working:
-
-1. Ensure exact title matching in [[WikiLinks]]
-2. Check for typos in relation types
-3. Verify both documents exist
-
-## Relations
-
-- implements [[Knowledge Format]] (How knowledge is structured)
-- relates_to [[Getting Started with Basic Memory]] (Setup and first steps)
-- relates_to [[Canvas]] (Creating visual knowledge maps)
-- relates_to [[CLI Reference]] (Command line tools)
-- enhanced_in_v0.13.0 [[OAuth Authentication Guide]] (Production authentication)
-- enhanced_in_v0.13.0 [[Project Management]] (Multi-project workflows)

docs/Welcome to Basic memory.md

@@ -1,68 +0,0 @@
----
-title: Introduction to Basic Memory
-type: docs
-permalink: docs/introduction
-tags:
-  - documentation
-  - index
-  - overview
----
-
-# BASIC MEMORY
-
-Basic Memory is a knowledge management system that allows you to build a persistent semantic graph from conversations
-with AI assistants. All knowledge is stored in standard Markdown files on your computer, giving you full control and
-ownership of your data.
-
-Basic Memory connects you and AI assistants through shared knowledge:
-
-1. **Captures knowledge** from natural conversations with AI assistants
-2. **Structures information** using simple semantic patterns in Markdown
-3. **Enables knowledge reuse** across different conversations and sessions
-4. **Maintains persistence** through local files you control completely
-
-Both you and AI assistants like Claude can read from and write to the same knowledge base, creating a continuous
-learning environment where each conversation builds upon previous ones.
-
-## Pick up your conversation right where you left off
-
-- AI assistants can load context from local files in a new conversation
-- Notes are saved locally as Markdown files in real time
-- No project knowledge or special prompting required
-
-![[Claude-Obsidian-Demo.mp4]]
-
-Basic Memory uses:
-
-- **Files as the source of truth** - Everything is stored in plain Markdown files
-- **Git-compatible storage** - All knowledge can be versioned, branched, and merged
-- **Local SQLite database** - For fast indexing and searching only (not primary storage)
-- **Model Context Protocol (MCP)** - For seamless AI assistant integration
-
-Basic Memory gives you complete control over your knowledge:
-
-- **Local-first storage** - All knowledge lives on your computer
-- **Standard file formats** - Plain Markdown compatible with any editor
-- **Directory organization** - Knowledge stored in `~/basic-memory` by default
-- **Version control ready** - Use git for history, branching, and collaboration
-- **Edit anywhere** - Modify files with any text editor or Obsidian
-
-Changes to files automatically sync with the knowledge graph, and AI assistants can see your edits in conversations.
-
-## Documentation Map
-
-Continue exploring Basic Memory with these guides:
-
-- Installation and setup [[Getting Started with Basic Memory]]
-- Comprehensive usage instructions [[User Guide]]
-- Detailed explanation of knowledge structure [[Knowledge Format]]
-- Obsidian integration guide [[Obsidian Integration]]
-- Canvas visualization guide [[Canvas]]
-- Command line tool reference [[CLI Reference]]
-- Reference for AI assistants using Basic Memory [[AI Assistant Guide]]
-- Technical implementation details [[Technical Information]]
-
-## Next Steps
-
-Start with the [[Getting Started with Basic Memory]] guide to install Basic Memory and configure it with your AI
-assistant.

docs/character-handling.md

@@ -0,0 +1,241 @@
+# 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

@@ -0,0 +1,821 @@
+# 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:
+- **Toggle cloud mode** - All regular `bm` commands work with cloud when enabled
+- **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 (when cloud mode is disabled)
+
+**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 --local-path ~/Documents/research
+bm project add work --local-path ~/work-notes
+bm project add temp  # 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. Enable Cloud Mode
+
+Authenticate and enable cloud mode:
+
+```bash
+bm cloud login
+```
+
+**What this does:**
+1. Opens browser to Basic Memory Cloud authentication page
+2. Stores authentication token in `~/.basic-memory/auth/token`
+3. **Enables cloud mode** - all CLI commands now work against cloud
+4. Validates your subscription status
+
+**Result:** All `bm project`, `bm tools` commands now work with cloud.
+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
+
+# Create cloud project WITH local sync
+bm project add research --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: `cloud_projects.research.local_path = "~/Documents/research"`
+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:
+- `Mode: Cloud (enabled)`
+- `Cloud instance is healthy`
+- Instructions for project sync commands
+
+## Working with Projects
+
+### Understanding Project Commands
+
+**Key concept:** When cloud mode is enabled, use regular `bm project` commands (not `bm cloud project`).
+
+```bash
+# In cloud mode:
+bm project list              # Lists cloud projects
+bm project add research      # Creates cloud project
+
+# In local mode:
+bm project list              # Lists local projects
+bm project add research ~/Documents/research  # Creates local project
+```
+
+### Creating Projects
+
+**Use case 1: Cloud-only project (no local sync)**
+
+```bash
+bm project add temp-notes
+```
+
+**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 --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:**
+- All projects in cloud (when cloud mode enabled)
+- Default project marked
+- Project paths shown
+
+**Future:** Will show sync status (synced/not synced, last sync time).
+
+## 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 Remote Files
+
+**Use case:** See what files exist on cloud without syncing.
+
+```bash
+# List all files in project
+bm project ls --name research
+
+# List files in subdirectory
+bm project ls --name research --path subfolder
+```
+
+**What happens:**
+1. Connects to cloud via rclone
+2. Lists files in remote project path
+3. No files transferred
+
+**Result:** See cloud file listing.
+
+## 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 --local-path ~/Documents/research
+bm project add work --local-path ~/work-notes
+bm project add personal --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 --local-path ~/Documents/research
+bm project add work --local-path ~/work
+
+# Cloud-only projects
+bm project add archive
+bm project add temp-notes
+
+# 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)
+
+Instead of toggling global cloud mode, you can route individual projects through the cloud using an API key. This lets you keep some projects local while others route through the 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. `BASIC_MEMORY_FORCE_LOCAL` env var
+3. Per-project cloud mode (API key)
+4. Global cloud mode (OAuth — deprecated fallback)
+5. Local ASGI transport (default)
+
+### 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.
+
+## Disable Cloud Mode
+
+Return to local mode (global):
+
+```bash
+bm cloud logout
+```
+
+**What this does:**
+1. Disables global cloud mode in config
+2. All commands now work locally (unless individual projects are set to cloud via `set-cloud`)
+3. Auth token remains (can re-enable with login)
+
+**Result:** All `bm` commands work with local projects again. Per-project cloud routing via API key continues to work independently of global cloud mode.
+
+## 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 Mode Management
+
+```bash
+bm cloud login              # Authenticate and enable global cloud mode (OAuth)
+bm cloud logout             # Disable global cloud mode
+bm cloud status             # Check cloud mode 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
+
+When cloud mode is enabled:
+
+```bash
+bm project list                           # List projects with mode column
+bm project add <name>                     # Create cloud project (no sync)
+bm project add <name> --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 remote files
+bm project ls --name <project>
+bm project ls --name <project> --path <subpath>
+```
+
+## Summary
+
+**Basic Memory Cloud uses project-scoped sync:**
+
+1. **Enable cloud mode** - `bm cloud login`
+2. **Install rclone** - `bm cloud setup`
+3. **Add projects with sync** - `bm project add research --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

@@ -0,0 +1,91 @@
+# 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

@@ -0,0 +1,130 @@
+# 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) ASCII / ANSI TUI Output
+
+Tools:
+- `search_notes(output_format="ascii" | "ansi")`
+- `read_note(output_format="ascii" | "ansi")`
+
+Expect:
+- ASCII table for search, header + content preview for note.
+- ANSI variants include color escape codes.
+
+Automated:
+- `uv run pytest test-int/mcp/test_output_format_ascii_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): __
+- ASCII/ANSI: __
+
+Decision + rationale: __

docs/publish.js

@@ -1,5 +0,0 @@
-var analyticsScript = document.createElement('script');
-analyticsScript.defer = true;
-analyticsScript.setAttribute('data-website-id', '8d51086e-5c67-401e-97b0-b24706a6d4f3');
-analyticsScript.src = 'https://cloud.umami.is/script.js';
-document.head.appendChild(analyticsScript);

docs/semantic-search.md

@@ -0,0 +1,265 @@
+# Semantic Search
+
+This guide covers Basic Memory's optional semantic (vector) search feature, which adds meaning-based retrieval alongside the existing full-text search.
+
+## Overview
+
+Basic Memory's default search uses full-text search (FTS) — keyword matching with boolean operators. 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 **opt-in** — existing behavior is completely unchanged unless you enable it. It works on both SQLite (local) and Postgres (cloud) backends.
+
+## Installation
+
+Semantic search dependencies (fastembed, sqlite-vec, openai) are **optional extras** — they are not installed with the base `basic-memory` package. Install them with:
+
+```bash
+pip install 'basic-memory[semantic]'
+```
+
+This keeps the base install lightweight and avoids platform-specific issues with ONNX Runtime wheels.
+
+### 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 FastEmbed provider uses ONNX Runtime, which dropped Intel Mac (x86_64) wheels starting in v1.24. Intel Mac users have two options:
+
+**Option 1: Use OpenAI embeddings (recommended)**
+
+Install only the OpenAI dependency manually — no ONNX Runtime or FastEmbed needed:
+
+```bash
+pip install openai sqlite-vec
+export BASIC_MEMORY_SEMANTIC_SEARCH_ENABLED=true
+export BASIC_MEMORY_SEMANTIC_EMBEDDING_PROVIDER=openai
+export OPENAI_API_KEY=sk-...
+```
+
+**Option 2: Pin an older ONNX Runtime**
+
+FastEmbed's ONNX Runtime dependency is unpinned, so you can constrain it to an older version that still ships Intel Mac wheels by passing both requirements in the same install command:
+
+```bash
+pip install 'basic-memory[semantic]' 'onnxruntime<1.24'
+```
+
+## Quick Start
+
+1. Install semantic extras:
+
+```bash
+pip install 'basic-memory[semantic]'
+```
+
+2. 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")
+
+# Traditional full-text search (still the default)
+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` | `false` | 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 semantic extras and enable
+pip install 'basic-memory[semantic]'
+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 reciprocal rank fusion (RRF). This is generally the best mode when you want both keyword precision and semantic recall.
+
+```python
+search_notes("authentication security", search_type="hybrid")
+```
+
+RRF merges the two ranked lists so that items appearing in both get a score boost, while items found by only one method still appear.
+
+### 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
+
+- **First enable**: After turning on `semantic_search_enabled` for the first time
+- **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 reciprocal rank fusion (RRF) to merge FTS and vector results:
+
+1. Run FTS search to get keyword-ranked results
+2. Run vector search to get similarity-ranked results
+3. For each result, compute: `score = 1/(k + fts_rank) + 1/(k + vector_rank)` where `k = 60`
+4. Sort by fused score
+
+Items found by both methods get a natural score boost. Items found by only one method still appear but rank lower.
+
+### 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-SCHEMA-IMPL.md

@@ -0,0 +1,365 @@
+# 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"
+
+
+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,
+) -> 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
+    """
+```
+
+### 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

@@ -0,0 +1,462 @@
+# 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` |
+
+### 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
+---
+
+# 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.
+
+### 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

@@ -0,0 +1,28 @@
+## 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).
+
+

examples/.obsidian/app.json

@@ -1,3 +0,0 @@
-{
-  "promptDelete": false
-}

examples/.obsidian/appearance.json

@@ -1 +0,0 @@
-{}

examples/.obsidian/community-plugins.json

@@ -1,3 +0,0 @@
-[
-  "optimize-canvas-connections"
-]

examples/.obsidian/core-plugins.json

@@ -1,31 +0,0 @@
-{
-  "file-explorer": true,
-  "global-search": true,
-  "switcher": true,
-  "graph": true,
-  "backlink": true,
-  "canvas": true,
-  "outgoing-link": true,
-  "tag-pane": true,
-  "properties": false,
-  "page-preview": true,
-  "daily-notes": true,
-  "templates": true,
-  "note-composer": true,
-  "command-palette": true,
-  "slash-command": false,
-  "editor-status": true,
-  "bookmarks": true,
-  "markdown-importer": false,
-  "zk-prefixer": false,
-  "random-note": false,
-  "outline": true,
-  "word-count": true,
-  "slides": false,
-  "audio-recorder": false,
-  "workspaces": false,
-  "file-recovery": true,
-  "publish": true,
-  "sync": true,
-  "webviewer": false
-}

examples/.obsidian/plugins/optimize-canvas-connections/main.js

@@ -1,158 +0,0 @@
-/*
-THIS IS A GENERATED/BUNDLED FILE BY ESBUILD
-if you want to view the source, please visit the github repository of this plugin
-*/
-
-var __defProp = Object.defineProperty;
-var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
-var __getOwnPropNames = Object.getOwnPropertyNames;
-var __hasOwnProp = Object.prototype.hasOwnProperty;
-var __export = (target, all) => {
-  for (var name in all)
-    __defProp(target, name, { get: all[name], enumerable: true });
-};
-var __copyProps = (to, from, except, desc) => {
-  if (from && typeof from === "object" || typeof from === "function") {
-    for (let key of __getOwnPropNames(from))
-      if (!__hasOwnProp.call(to, key) && key !== except)
-        __defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
-  }
-  return to;
-};
-var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
-
-// main.ts
-var main_exports = {};
-__export(main_exports, {
-  default: () => OptimizeCanvasConnectionsPlugin
-});
-module.exports = __toCommonJS(main_exports);
-var import_obsidian = require("obsidian");
-var OptimizeCanvasConnectionsPlugin = class extends import_obsidian.Plugin {
-  async onload() {
-    this.addCommand({
-      id: "optimize-preserve-axes-selection",
-      name: "Optimize selection (preserve axes)",
-      checkCallback: (checking) => {
-        const canvasView = app.workspace.getActiveViewOfType(import_obsidian.ItemView);
-        if ((canvasView == null ? void 0 : canvasView.getViewType()) == "canvas") {
-          if (!checking) {
-            this.optimize("preserve-axes");
-          }
-          return true;
-        }
-        return false;
-      }
-    });
-    this.addCommand({
-      id: "optimize-shortest-path-selection",
-      name: "Optimize selection (shortest path)",
-      checkCallback: (checking) => {
-        const canvasView = app.workspace.getActiveViewOfType(import_obsidian.ItemView);
-        if ((canvasView == null ? void 0 : canvasView.getViewType()) == "canvas") {
-          if (!checking) {
-            this.optimize("shortest-path");
-          }
-          return true;
-        }
-        return false;
-      }
-    });
-  }
-  onunload() {
-  }
-  async optimize(option) {
-    const canvasView = app.workspace.getActiveViewOfType(import_obsidian.ItemView);
-    const canvas = canvasView == null ? void 0 : canvasView.canvas;
-    const currentSelection = canvas == null ? void 0 : canvas.selection;
-    let selectedIDs = new Array();
-    currentSelection.forEach(function(selection) {
-      selectedIDs.push(selection.id);
-    });
-    let applyToAll = false;
-    if (selectedIDs.length == 0) {
-      applyToAll = true;
-    }
-    for (let [edgeKey, edge] of canvas["edges"]) {
-      let fromNode = edge["from"]["node"];
-      let toNode = edge["to"]["node"];
-      let fromPossibilities = [edge["from"]["side"]];
-      if (applyToAll || selectedIDs.includes(fromNode["id"])) {
-        switch (option) {
-          case "shortest-path":
-            fromPossibilities = ["top", "bottom", "left", "right"];
-            break;
-          case "preserve-axes":
-            switch (edge["from"]["side"]) {
-              case "top":
-              case "bottom":
-                fromPossibilities = ["top", "bottom"];
-                break;
-              case "left":
-              case "right":
-                fromPossibilities = ["left", "right"];
-                break;
-            }
-        }
-      }
-      let toPossibilities = [edge["to"]["side"]];
-      if (applyToAll || selectedIDs.includes(toNode["id"])) {
-        switch (option) {
-          case "shortest-path":
-            toPossibilities = ["top", "bottom", "left", "right"];
-            break;
-          case "preserve-axes":
-            switch (edge["to"]["side"]) {
-              case "top":
-              case "bottom":
-                toPossibilities = ["top", "bottom"];
-                break;
-              case "left":
-              case "right":
-                toPossibilities = ["left", "right"];
-                break;
-            }
-        }
-      }
-      let distances = [];
-      for (const fromSide of fromPossibilities) {
-        let fromPoint = { "x": 0, "y": 0 };
-        if (fromSide == "top") {
-          fromPoint = { "x": fromNode["x"] + fromNode["width"] / 2, "y": fromNode["y"] };
-        } else if (fromSide == "bottom") {
-          fromPoint = { "x": fromNode["x"] + fromNode["width"] / 2, "y": fromNode["y"] + fromNode["height"] };
-        } else if (fromSide == "left") {
-          fromPoint = { "x": fromNode["x"], "y": fromNode["y"] + fromNode["height"] / 2 };
-        } else if (fromSide == "right") {
-          fromPoint = { "x": fromNode["x"] + fromNode["width"], "y": fromNode["y"] + fromNode["height"] / 2 };
-        }
-        for (const toSide of toPossibilities) {
-          let toPoint = { "x": 0, "y": 0 };
-          if (toSide == "top") {
-            toPoint = { "x": toNode["x"] + toNode["width"] / 2, "y": toNode["y"] };
-          } else if (toSide == "bottom") {
-            toPoint = { "x": toNode["x"] + toNode["width"] / 2, "y": toNode["y"] + toNode["height"] };
-          } else if (toSide == "left") {
-            toPoint = { "x": toNode["x"], "y": toNode["y"] + toNode["height"] / 2 };
-          } else if (toSide == "right") {
-            toPoint = { "x": toNode["x"] + toNode["width"], "y": toNode["y"] + toNode["height"] / 2 };
-          }
-          distances.push({
-            "fromSide": fromSide,
-            "toSide": toSide,
-            "distance": (toPoint.x - fromPoint.x) ** 2 + (toPoint.y - fromPoint.y) ** 2
-          });
-        }
-      }
-      distances = distances.sort(function(a, b) {
-        return a.distance - b.distance;
-      });
-      edge["from"]["side"] = distances[0]["fromSide"];
-      edge["to"]["side"] = distances[0]["toSide"];
-      edge.render();
-    }
-    canvas.requestSave();
-  }
-};
-
-/* nosourcemap */

examples/.obsidian/plugins/optimize-canvas-connections/manifest.json

@@ -1,10 +0,0 @@
-{
-	"id": "optimize-canvas-connections",
-	"name": "Optimize Canvas Connections",
-	"version": "1.0.0",
-	"minAppVersion": "1.1.9",
-	"description": "An Obsidian plugin that declutters a canvas by reconnecting notes using their nearest edges.",
-	"author": "Félix Chénier",
-	"authorUrl": "https://felixchenier.uqam.ca",
-	"isDesktopOnly": false
-}

examples/.obsidian/publish.json

@@ -1,6 +0,0 @@
-{
-  "siteId": null,
-  "host": null,
-  "included": [],
-  "excluded": []
-}

examples/Coffee Notes/Brewing Equipment.md

@@ -1,83 +0,0 @@
----
-title: Brewing Equipment
-type: note
-permalink: coffee/brewing-equipment
-tags:
-- '#coffee'
-- '#equipment'
-- '#gear'
-- '#brewing'
-- '#demo'
----
-
-# Brewing Equipment
-
-Essential tools and equipment for brewing coffee, their characteristics, and how they affect the brewing process.
-
-## Overview
-
-The equipment used to brew coffee plays a crucial role in determining the final cup quality. From grinders to brewers to kettles, each piece of equipment contributes to different aspects of the brewing process.
-
-## Observations
-
-- [principle] Equipment quality often has a bigger impact on consistency than on absolute quality potential #quality
-- [principle] Good grind consistency is the most important technical factor in extraction quality #grind
-- [investment] A good burr grinder is often the most important investment for improving home coffee #gear
-- [technique] Equipment maintenance and cleaning significantly impact flavor consistency over time #maintenance
-
-## Grinders
-
-- [equipment] Burr grinders crush beans between two abrasive surfaces for more consistent particle size #grinders
-- [equipment] Blade grinders chop beans unevenly, leading to inconsistent extraction #grinders
-- [equipment] Flat burr grinders produce very consistent particle size but generate more heat #burrs
-- [equipment] Conical burr grinders create slightly less uniform grounds but with less heat and noise #burrs
-- [feature] Grind adjustment mechanisms range from stepped to stepless for different precision levels #adjustment
-- [feature] Retention (grounds trapped in grinder) affects dose consistency and freshness #retention
-- [price] Hand grinders offer excellent value, with models like Timemore C2 and 1Zpresso JX providing excellent results around $100-150 #budget
-- [price] Entry-level electric burr grinders like Baratza Encore start around $170 but provide significant improvement over blade grinders #value
-
-## Brewers
-
-### Pour Over Brewers
-- [equipment] Hario V60 uses a conical design with spiral ridges to control flow rate #pourover
-- [equipment] Kalita Wave has a flat bottom with three small holes for more consistent extraction #pourover
-- [equipment] Chemex combines brewer and server with thick proprietary filters for ultra-clean cup #pourover
-- [material] Ceramic brewers retain heat better than plastic but are more fragile #materials
-- [material] Glass brewers provide neutral flavor but less heat retention #materials
-- [material] Plastic brewers are inexpensive, durable, and surprisingly good for heat retention #materials
-
-### Immersion Brewers
-- [equipment] French Press uses a metal mesh to separate grounds, allowing oils and fine particles to pass #immersion
-- [equipment] AeroPress uses pressure and paper filter for clean, versatile brewing #immersion
-- [equipment] Clever Dripper combines immersion and drip methods with a valve mechanism #hybrid
-- [material] Glass French presses look elegant but break easily and have poor heat retention #materials
-- [material] Stainless steel or ceramic French presses offer better durability and heat retention #materials
-
-### Pressure Brewers
-- [equipment] Espresso machines use 9 bars of pressure, requiring significant investment for good results #espresso
-- [equipment] Moka pot uses steam pressure for strong, concentrated coffee at affordable price #moka
-- [equipment] Manual lever machines like Flair or Robot provide espresso-style coffee with manual control #manual_espresso
-
-## Kettles
-
-- [equipment] Gooseneck kettles provide precision pouring control essential for pour over methods #kettles
-- [feature] Variable temperature kettles allow precise temperature control for different roast levels #temp_control
-- [feature] Flow restrictors can help beginners maintain consistent pour rates #pour_control
-- [material] Electric kettles offer convenience and temperature stability #convenience
-- [material] Stovetop kettles may be more durable but offer less temperature control #durability
-
-## Accessories
-
-- [equipment] Coffee scale with 0.1g precision helps maintain consistent ratios #measurement
-- [equipment] Timer ensures consistent extraction times #consistency
-- [equipment] Quality filters significantly impact flavor clarity and body #filters
-- [equipment] Storage containers with one-way valves help preserve bean freshness #storage
-- [equipment] Blind shaker or dosing cup reduces grinder mess and improves workflow #workflow
-
-## Relations
-
-- improves [[Coffee Brewing Methods]]
-- affects [[Flavor Extraction]]
-- requires [[Proper Maintenance]]
-- enhances [[Home Coffee Setup]]
-- part_of [[Coffee Knowledge Base]]

examples/Coffee Notes/Coffee Bean Origins.md

@@ -1,78 +0,0 @@
----
-title: Coffee Bean Origins
-type: note
-permalink: coffee/coffee-bean-origins
-tags:
-- '#coffee'
-- '#origins'
-- '#beans'
-- '#regions'
-- '#demo'
----
-
-# Coffee Bean Origins
-
-An exploration of coffee-growing regions around the world and how geography, climate, and processing methods affect flavor profiles.
-
-## Overview
-
-Coffee beans are grown in various regions around the world, primarily in what's known as the "Coffee Belt" - the area between the Tropics of Cancer and Capricorn. The flavor characteristics of coffee beans are influenced by:
-
-- Geographic region and climate
-- Altitude
-- Soil composition
-- Variety of coffee plant
-- Processing method
-- Harvest and sorting practices
-
-## Observations
-
-- [principle] Higher altitude generally produces harder, denser beans with more complex acidity #altitude
-- [region] Ethiopian beans often feature bright, fruity notes with floral aromatics #ethiopia
-- [region] Colombian coffee typically offers balanced acidity with caramel sweetness and nutty undertones #colombia
-- [region] Guatemalan coffee presents complex acidity with chocolate notes and sometimes spice characteristics #guatemala
-- [region] Brazilian coffee tends toward nutty, chocolate notes with lower acidity and fuller body #brazil
-- [region] Kenyan coffee is known for bright, wine-like acidity and berry or citrus notes #kenya
-- [processing] Natural (dry) processing tends to create fruitier, more fermented flavors #processing
-- [processing] Washed (wet) processing generally results in cleaner, brighter cups with more clarity #processing
-- [processing] Honey processing creates a middle ground with some fruity notes while maintaining clarity #processing
-- [factor] Shade-grown coffee typically develops more slowly, resulting in more complex flavors #cultivation
-- [factor] Soil volcanic soil often imparts distinctive mineral characteristics to coffee #terroir
-- [variety] Gesha/Geisha variety is known for exceptional floral and tea-like qualities #varieties
-- [variety] Bourbon varieties often feature sweet, complex cup profiles #varieties
-- [variety] Robusta beans have higher caffeine content but generally less complex flavor than Arabica #varieties
-
-## Major Growing Regions
-
-- [africa] Ethiopian coffees: Yirgacheffe, Sidamo, Harrar regions each with distinctive profiles #ethiopia
-- [africa] Kenyan coffees: Often categorized by grade (AA, AB, etc.) based on bean size #kenya
-- [americas] Colombian regions: Huila, Nariño, Antioquia each with unique characteristics #colombia
-- [americas] Central American producers: Guatemala, Costa Rica, Panama known for balanced profiles #central_america
-- [americas] Brazilian regions: Cerrado, Sul de Minas, Mogiana with varying profiles #brazil
-- [asia] Indonesian islands: Sumatra, Java, Sulawesi producing earthy, full-bodied coffees #indonesia
-- [asia] Vietnamese coffee: World's largest Robusta producer, often used in blends and commercial coffee #vietnam
-
-## Processing Methods
-
-- [natural] Beans dried inside the fruit, creating fruity, fermented notes and heavier body #processing
-- [washed] Fruit removed before drying, resulting in cleaner cup with more pronounced acidity #processing
-- [honey] Some fruit mucilage left on during drying, creates balanced sweetness and body #processing
-- [wet-hulled] Unique to Indonesia, creates earthy, herbal, low-acid profiles #processing
-- [experimental] Anaerobic fermentation, wine-yeast inoculation, and other newer methods #innovation
-
-## Tasting Notes by Region
-
-- [ethiopia] Blueberry, jasmine, bergamot, stone fruit, citrus #flavor_notes
-- [kenya] Blackcurrant, tomato, tropical fruit, wine-like acidity #flavor_notes
-- [colombia] Caramel, nuts, red apple, chocolate, balanced acidity #flavor_notes
-- [guatemala] Chocolate, spice, green apple, balanced #flavor_notes
-- [brazil] Nuts, chocolate, low acidity, full body #flavor_notes
-- [indonesia] Earthy, herbal, spice, cedar, full body, low acidity #flavor_notes
-
-## Relations
-
-- influences [[Flavor Extraction]]
-- pairs_with [[Coffee Brewing Methods]]
-- affects [[Tasting Notes]]
-- relates_to [[Specialty Coffee]]
-- part_of [[Coffee Knowledge Base]]

examples/Coffee Notes/Coffee Brewing Methods.md

@@ -1,70 +0,0 @@
----
-title: Coffee Brewing Methods
-type: note
-permalink: coffee/coffee-brewing-methods
-tags:
-- '#coffee'
-- '#brewing'
-- '#methods'
-- '#demo'
----
-
-# Coffee Brewing Methods
-
-An exploration of different coffee brewing techniques, their characteristics, and how they affect flavor extraction.
-
-## Overview
-
-Coffee brewing is both an art and a science. Different brewing methods extract different compounds from coffee beans, resulting in unique flavor profiles, body, and mouthfeel. The key variables in any brewing method are:
-
-- Grind size
-- Water temperature
-- Brew time
-- Coffee-to-water ratio
-- Agitation/turbulence
-
-## Observations
-
-- [principle] Coffee extraction follows a predictable pattern: acids extract first, then sugars, then bitter compounds #extraction
-- [method] Pour over methods generally produce cleaner, brighter cups with more distinct flavor notes #clarity
-- [method] Immersion methods like French press create fuller body and more rounded flavors #body
-- [technique] Water at 195-205°F (90-96°C) extracts optimal flavor compounds for most brewing methods #temperature
-- [technique] Grind size directly correlates with ideal extraction time (finer = shorter, coarser = longer) #grind
-- [preference] Medium-light roasts often showcase more origin characteristics in pour over methods #roast
-- [equipment] Burr grinders produce more consistent particle size than blade grinders, resulting in more even extraction #gear
-- [ratio] 1:15 to 1:17 coffee-to-water ratio (by weight) works well for most brew methods #brewing
-- [science] Different brewing temperatures extract different chemical compounds from the beans #chemistry
-- [technique] Bloom phase (pre-infusion with small amount of water) allows CO2 to escape and improves extraction #bloom
-
-## Pour Over Methods
-
-- [method] V60 produces very clean cup with excellent clarity of flavor #pourover
-- [method] Chemex uses thicker filter paper, resulting in even cleaner cup with fewer oils #pourover
-- [method] Kalita Wave provides more consistent extraction due to flat bottom design #pourover
-- [technique] Concentric circular pouring pattern ensures even saturation of grounds #technique
-- [timing] Most pour over methods complete in 2:30-3:30 total brew time #brewing
-
-## Immersion Methods
-
-- [method] French Press creates full-bodied cup with rich mouthfeel due to metal filter allowing oils to pass #immersion
-- [method] AeroPress is versatile, capable of producing both espresso-like and filter-style coffee #immersion
-- [method] Cold brew uses time instead of heat to extract, resulting in lower acidity #immersion
-- [technique] French press ideal steep time is 4-5 minutes before plunging #timing
-- [technique] AeroPress inverted method prevents dripping during extraction phase #technique
-
-## Pressure Methods
-
-- [method] Espresso uses 9 bars of pressure to force water through finely ground coffee #pressure
-- [method] Moka pot uses steam pressure to push water through grounds, creating strong, concentrated coffee #pressure
-- [technique] Espresso requires very fine grind, almost powder-like consistency #grind
-- [timing] Espresso shots typically extract in 25-30 seconds #timing
-- [principle] Pressure methods can extract compounds that aren't soluble in regular brewing methods #extraction
-
-## Relations
-
-- requires [[Proper Grinding Technique]]
-- affects [[Flavor Extraction]]
-- pairs_with [[Coffee Bean Origins]]
-- uses [[Brewing Equipment]]
-- influences [[Tasting Notes]]
-- part_of [[Coffee Knowledge Base]]

examples/Coffee Notes/Coffee Flavor Map.md

@@ -1,89 +0,0 @@
----
-title: Coffee Flavor Map
-type: note
-permalink: coffee/coffee-flavor-map
-tags:
-- '#coffee'
-- '#visualization'
-- '#canvas'
-- '#demo'
----
-
-# Coffee Flavor Map
-
-A visual mapping of coffee flavor attributes, brewing methods, and their relationships. This note describes a canvas visualization that could be generated to demonstrate Basic Memory's visualization capabilities.
-
-## Overview
-
-The Coffee Flavor Map provides a visual representation of how different brewing methods, coffee origins, and equipment choices affect flavor outcomes. This canvas visualization helps users understand the complex relationships in coffee brewing and tasting.
-
-## Canvas Visualization Elements
-
-### Core Nodes
-- **Flavor Attributes**: Acidity, Sweetness, Body, Clarity, Bitterness, Complexity
-- **Brewing Methods**: Pour Over, French Press, AeroPress, Espresso, Moka Pot, Cold Brew
-- **Origin Regions**: Ethiopia, Kenya, Colombia, Brazil, Guatemala, Indonesia
-- **Equipment Elements**: Grinder Quality, Water Temperature, Brewing Device, Filter Type
-
-### Node Connections
-- Lines connecting brewing methods to their typical flavor outcomes
-- Arrows showing how equipment choices affect extraction variables
-- Connections between origins and their characteristic flavor profiles
-- Highlighting of optimal brewing methods for different origins
-
-### Visual Organization
-- Flavor outcomes in the center
-- Brewing methods on the left side
-- Origins on the right side
-- Equipment variables at the bottom
-- Color coding by category (methods, origins, equipment, flavors)
-
-## Using This Visualization
-
-### For Coffee Exploration
-- Identify which brewing methods might highlight the characteristics you prefer
-- See which origins naturally pair well with your preferred brewing method
-- Understand how equipment changes can modify flavor outcomes
-- Visualize the complex interplay between all coffee variables
-
-### As a Basic Memory Demo
-- Demonstrates Canvas visualization capabilities
-- Shows how relations can be visually mapped
-- Illustrates complex knowledge organization
-- Provides an intuitive way to navigate coffee knowledge
-
-## How To Generate This Canvas
-
-In a conversation with Claude, you could request:
-
-```
-Please create a canvas visualization mapping the relationships between coffee brewing methods, origins, and flavor outcomes. Show how different equipment and techniques influence extraction and resulting flavor profiles.
-```
-
-This would generate a `.canvas` file in your Basic Memory directory that could be opened with Obsidian for an interactive visualization of these coffee relationships.
-
-## Example Visualization Snippets
-
-### Pour Over Method Node
-- Connected to: High Clarity, Bright Acidity, Medium Body
-- Best pairs with: Ethiopian and Kenyan beans
-- Equipment dependencies: Gooseneck Kettle, Paper Filter, Burr Grinder
-
-### Ethiopian Coffee Node
-- Characteristic flavors: Floral, Fruity, Bright
-- Best brewing methods: Pour Over, AeroPress
-- Challenging with: French Press (loses clarity of delicate notes)
-
-### Grind Size Node
-- Affects: Extraction Rate, Flavor Balance
-- Fine grind increases: Extraction Speed, Surface Area
-- Coarse grind increases: Flow Rate, Reduces Bitter Compounds
-
-## Relations
-
-- visualizes [[Coffee Knowledge Base]]
-- relates_to [[Coffee Brewing Methods]]
-- relates_to [[Coffee Bean Origins]]
-- relates_to [[Flavor Extraction]]
-- relates_to [[Tasting Notes]]
-- demonstrates [[Canvas]]

examples/Coffee Notes/Coffee Knowledge Base.md

@@ -1,73 +0,0 @@
----
-title: Coffee Knowledge Base
-type: note
-permalink: coffee/coffee-knowledge-base
-tags:
-- '#coffee'
-- '#index'
-- '#demo'
-- '#knowledge'
----
-
-# Coffee Knowledge Base
-
-A comprehensive collection of coffee knowledge, from bean origins to brewing methods to tasting notes. This knowledge base demonstrates Basic Memory's ability to organize and connect information in a meaningful way.
-
-## Overview
-
-This Coffee Knowledge Base captures key information about coffee, structured with semantic observations and relations that connect different aspects of coffee knowledge. It serves as both a useful reference for coffee enthusiasts and a demonstration of how Basic Memory organizes information.
-
-## Key Topics
-
-### Core Coffee Knowledge
-
-- [[Coffee Brewing Methods]] - Different techniques for preparing coffee
-- [[Coffee Bean Origins]] - Where coffee comes from and how region affects flavor
-- [[Brewing Equipment]] - Tools and devices used to prepare coffee
-- [[Flavor Extraction]] - The science of dissolving flavor compounds from coffee
-- [[Tasting Notes]] - How to taste and describe coffee flavors
-
-### Brewing Techniques
-
-- Proper grinding is fundamental to good extraction
-- Water quality significantly impacts flavor
-- Different brewing methods highlight different characteristics
-- Time, temperature, and grind size are the key variables to control
-- Freshness of beans dramatically affects quality
-
-### Coffee Preferences
-
-- Light roasts preserve more origin characteristics and acidity
-- Dark roasts emphasize body and chocolatey/roasted flavors
-- Pour over methods highlight clarity and distinct flavor notes
-- Immersion methods create fuller body and rounded flavor
-- Personal preference matters more than "correctness"
-
-## Using This Knowledge Base
-
-### For Learning
-
-Use this knowledge base to:
-- Understand coffee fundamentals
-- Explore connections between brewing methods and flavor outcomes
-- Learn how different origins produce distinct flavor profiles
-- Discover how equipment affects the brewing process
-- Develop a vocabulary for describing coffee experiences
-
-### As a Demo
-
-This knowledge base demonstrates:
-- Semantic knowledge organization with categories and relations
-- Building connections between related concepts
-- Creating a navigable knowledge graph
-- Structuring information in a way both humans and AI assistants can understand
-- How Basic Memory enables persistent knowledge across conversations
-
-## Relations
-
-- contains [[Coffee Brewing Methods]]
-- contains [[Coffee Bean Origins]]
-- contains [[Brewing Equipment]]
-- contains [[Flavor Extraction]]
-- contains [[Tasting Notes]]
-- demonstrates [[Basic Memory Capabilities]]

examples/Coffee Notes/Flavor Extraction.md

@@ -1,79 +0,0 @@
----
-title: Flavor Extraction
-type: note
-permalink: coffee/flavor-extraction
-tags:
-- '#coffee'
-- '#extraction'
-- '#brewing'
-- '#science'
-- '#demo'
----
-
-# Flavor Extraction
-
-Understanding the science of coffee extraction, how different compounds dissolve at different rates, and how to control extraction to achieve desired flavor profiles.
-
-## Overview
-
-Coffee extraction is the process of dissolving flavor compounds from ground coffee into water. The science of extraction is key to producing a balanced, flavorful cup. Extraction is affected by numerous variables including grind size, water temperature, contact time, agitation, and pressure.
-
-## Observations
-
-- [science] Coffee contains over 1,000 aroma compounds and hundreds of flavor compounds #chemistry
-- [principle] Extraction occurs in a predictable sequence: acids → sugars → bitter compounds #extraction_order
-- [principle] Under-extraction results in sour, bright, thin coffee lacking sweetness and body #under_extraction
-- [principle] Over-extraction results in bitter, hollow, astringent flavors #over_extraction
-- [principle] The goal is typically balanced extraction (18-22% of coffee solubles dissolved) #balanced_extraction
-- [technique] Finer grind size increases extraction rate due to greater surface area #grind_size
-- [technique] Higher water temperature increases extraction rate and solubility of compounds #temperature
-- [technique] Longer contact time allows more complete extraction #brew_time
-- [technique] Agitation (stirring, turbulence) increases extraction rate by preventing saturation zones #agitation
-- [technique] Pressure (as in espresso) can extract compounds that aren't water-soluble at atmospheric pressure #pressure
-
-## Factors Affecting Extraction
-
-- [factor] Grind size: Finer = faster extraction, coarser = slower extraction #grind
-- [factor] Water temperature: Higher = faster extraction, lower = slower extraction #temperature
-- [factor] Contact time: Longer = more extraction, shorter = less extraction #time
-- [factor] Agitation: More = faster extraction, less = slower extraction #agitation
-- [factor] Coffee-to-water ratio: More coffee = lower extraction percentage #ratio
-- [factor] Water quality: Mineral content affects extraction of different compounds #water
-- [factor] Roast level: Darker roasts extract more easily than lighter roasts #roast
-- [factor] Bean density: Denser beans (typically high-altitude) require more effort to extract #density
-- [factor] Freshness: Freshly roasted coffee extracts differently than aged coffee #freshness
-- [factor] Brewing method: Different methods extract different compounds at different rates #method
-
-## Signs of Extraction Levels
-
-- [under] Sour, bright, lack of sweetness, thin body, quick finish #flavor
-- [under] Typically from: too coarse grind, too cool water, too short brew time #causes
-- [balanced] Sweet, bright but not sour, rich but not bitter, pleasing finish #flavor
-- [balanced] Achieved through proper ratio of variables for given coffee #technique
-- [over] Bitter, hollow, astringent, dry finish, sometimes papery #flavor
-- [over] Typically from: too fine grind, too hot water, too long brew time #causes
-
-## Measuring Extraction
-
-- [method] Total Dissolved Solids (TDS) meters measure concentration of coffee solution #measurement
-- [method] Extraction yield = percentage of coffee grounds dissolved in the final brew #calculation
-- [preference] Specialty coffee typically targets 18-22% extraction yield #standards
-- [preference] Some specialty light roasts may taste best at higher extraction percentages #speciality
-
-## Controlling Extraction
-
-- [technique] Adjust grind size as primary extraction control #basics
-- [technique] Use water temperature to fine-tune extraction #fine_tuning
-- [technique] Modify pour technique to control agitation level #technique
-- [technique] Adjust coffee-to-water ratio to balance strength and extraction #ratio
-- [technique] Pre-infusion (blooming) helps achieve even extraction #blooming
-- [technique] Pulse pouring creates different extraction dynamics than continuous pour #pour_technique
-
-## Relations
-
-- affected_by [[Coffee Brewing Methods]]
-- influenced_by [[Coffee Bean Origins]]
-- enhanced_by [[Brewing Equipment]]
-- determines [[Tasting Notes]]
-- requires [[Water Quality]]
-- part_of [[Coffee Knowledge Base]]

examples/Coffee Notes/Perfect Pour Over Coffee Method.canvas

@@ -1,161 +0,0 @@
-{
-	"nodes":[
-		{
-			"id":"node-5",
-			"type":"text",
-			"text":"## Main Pour Phase\n- Use concentric circles from center outward\n- Maintain steady, controlled flow rate\n- Avoid pouring directly on filter walls\n- Keep water level consistent\n- Pulse pour in 2-3 stages (or continuous pour)\n- Total brew time target: 2:30-3:30",
-			"position":{"x":450,"y":200},
-			"x":530,
-			"y":-100,
-			"width":300,
-			"height":200,
-			"color":"1"
-		},
-		{
-			"id":"node-8",
-			"type":"text",
-			"text":"## Drawdown\n- Allow water to fully drain\n- Flat bed indicates even extraction\n- Total brew time should be ~2:30-3:30\n- Remove filter promptly after brewing",
-			"position":{"x":450,"y":700},
-			"x":540,
-			"y":375,
-			"width":300,
-			"height":150,
-			"color":"1"
-		},
-		{
-			"id":"node-6",
-			"type":"text",
-			"text":"## Pour Pattern\n\nConcentric circles ensure even saturation of coffee grounds. Begin at the center and work outward, avoiding filter edges. Pour height of 1-2 inches above coffee bed.",
-			"position":{"x":250,"y":450},
-			"x":960,
-			"y":25,
-			"width":300,
-			"height":150,
-			"color":"5"
-		},
-		{
-			"id":"node-12",
-			"type":"text",
-			"text":"## Tasting Notes\n\n- Balanced extraction: sweet, bright, complex\n- Under-extraction: sour, lacking sweetness\n- Over-extraction: bitter, astringent, hollow\n\nTake notes on each brew to track improvements and preferences.",
-			"position":{"x":-250,"y":700},
-			"x":1020,
-			"y":420,
-			"width":300,
-			"height":150,
-			"color":"6"
-		},
-		{
-			"id":"node-9",
-			"type":"text",
-			"text":"## Troubleshooting\n\n- Too sour/weak: Grind finer, water hotter, increase brew time\n- Too bitter/strong: Grind coarser, water cooler, decrease brew time\n- Uneven extraction: Improve pour technique, better grinder\n- Channeling: More careful pouring, better bloom\n- Slow drawdown: Coarser grind, less agitation\n- Fast drawdown: Finer grind, more careful pouring",
-			"position":{"x":100,"y":700},
-			"x":30,
-			"y":570,
-			"width":300,
-			"height":200,
-			"color":"6"
-		},
-		{
-			"id":"node-3",
-			"type":"text",
-			"text":"## Preparation\n- Heat water to 195-205°F (90-96°C)\n- Measure coffee (1:15 to 1:17 ratio)\n- Medium-fine grind (sea salt consistency)\n- Rinse filter with hot water\n- Discard rinse water\n- Add ground coffee to filter\n- Level coffee bed",
-			"position":{"x":-250,"y":200},
-			"x":30,
-			"y":-500,
-			"width":300,
-			"height":200,
-			"color":"3"
-		},
-		{
-			"id":"node-1",
-			"type":"text",
-			"text":"# Perfect Pour Over Method\n\nA systematic approach to brewing exceptional pour over coffee by controlling key variables and following proper technique.",
-			"position":{"x":0,"y":0},
-			"x":-580,
-			"y":-760,
-			"width":400,
-			"height":120,
-			"color":"4"
-		},
-		{
-			"id":"node-10",
-			"type":"text",
-			"text":"## Grinding Parameters\n\n- V60: Medium-fine (sea salt)\n- Chemex: Medium (slightly coarser than V60)\n- Kalita Wave: Medium (between V60 and Chemex)\n\nConsistent particle size is critical; use quality burr grinder.",
-			"position":{"x":-250,"y":450},
-			"x":30,
-			"y":-910,
-			"width":300,
-			"height":150,
-			"color":"5"
-		},
-		{
-			"id":"node-2",
-			"type":"text",
-			"text":"## Equipment Setup\n- Clean V60/Chemex/Kalita Wave\n- Paper filter (rinsed)\n- Server/mug\n- Scale with timer\n- Gooseneck kettle\n- Burr grinder\n- Fresh coffee beans",
-			"position":{"x":-600,"y":200},
-			"x":-530,
-			"y":-500,
-			"width":300,
-			"height":200,
-			"color":"3"
-		},
-		{
-			"id":"node-4",
-			"type":"text",
-			"text":"## The Bloom\n- Start timer\n- Pour 2-3x coffee weight water\n- Ensure all grounds are saturated\n- Gentle stir or swirl if needed\n- Allow 30-45 seconds for degassing\n- Look for bubbling and dome formation",
-			"position":{"x":100,"y":200},
-			"x":530,
-			"y":-500,
-			"width":300,
-			"height":200,
-			"color":"1"
-		},
-		{
-			"id":"node-13",
-			"type":"text",
-			"text":"## Coffee-to-Water Ratio\n\n- Standard: 1:15 to 1:17 (coffee:water)\n- Stronger cup: 1:15 (67g/L)\n- Medium cup: 1:16 (62.5g/L)\n- Lighter cup: 1:17 (58.8g/L)\n\nExample: For 300ml water, use ~18-20g coffee",
-			"position":{"x":-600,"y":700},
-			"x":30,
-			"y":-100,
-			"width":300,
-			"height":150,
-			"color":"6"
-		},
-		{
-			"id":"node-7",
-			"type":"text",
-			"text":"## Brew Time Guideline\n\n- Bloom: 30-45 seconds\n- First pour: 1:00-1:15\n- Second pour: 1:45-2:00\n- Final pour: 2:15-2:30\n- Drawdown complete: 2:45-3:30\n\nAdjust for taste: shorter for lighter, longer for stronger",
-			"position":{"x":600,"y":450},
-			"x":-80,
-			"y":220,
-			"width":300,
-			"height":200,
-			"color":"5"
-		},
-		{
-			"id":"node-11",
-			"type":"text",
-			"text":"## Water Quality\n\n- Clean, filtered water\n- No strong odors or flavors\n- Ideal TDS: 75-150 ppm\n- Ideal pH: 7.0-7.5\n- Avoid distilled water (lacks minerals)\n- Avoid hard water (scaling issues)",
-			"position":{"x":-600,"y":450},
-			"x":-780,
-			"y":-125,
-			"width":300,
-			"height":150,
-			"color":"5"
-		}
-	],
-	"edges":[
-		{"id":"edge-1","fromNode":"node-1","fromSide":"bottom","toNode":"node-2","toSide":"top","label":"Step 1"},
-		{"id":"edge-2","fromNode":"node-2","fromSide":"right","toNode":"node-3","toSide":"left","label":"Step 2"},
-		{"id":"edge-3","fromNode":"node-3","fromSide":"right","toNode":"node-4","toSide":"left","label":"Step 3"},
-		{"id":"edge-4","fromNode":"node-4","fromSide":"bottom","toNode":"node-5","toSide":"top","label":"Step 4"},
-		{"id":"edge-5","fromNode":"node-5","fromSide":"bottom","toNode":"node-8","toSide":"top","label":"Step 5"},
-		{"id":"edge-6","fromNode":"node-5","fromSide":"left","toNode":"node-7","toSide":"right","label":"Timing"},
-		{"id":"edge-7","fromNode":"node-5","fromSide":"right","toNode":"node-6","toSide":"left","label":"Technique"},
-		{"id":"edge-8","fromNode":"node-8","fromSide":"left","toNode":"node-9","toSide":"right","label":"if problems"},
-		{"id":"edge-9","fromNode":"node-3","fromSide":"top","toNode":"node-10","toSide":"bottom","label":"Grinding details"},
-		{"id":"edge-10","fromNode":"node-2","fromSide":"bottom","toNode":"node-11","toSide":"right","label":"Water details"},
-		{"id":"edge-11","fromNode":"node-8","fromSide":"right","toNode":"node-12","toSide":"left","label":"Evaluate"},
-		{"id":"edge-12","fromNode":"node-3","fromSide":"bottom","toNode":"node-13","toSide":"top","label":"Ratio details"}
-	]
-}

examples/Coffee Notes/Tasting Notes.md

@@ -1,84 +0,0 @@
----
-title: Tasting Notes
-type: note
-permalink: coffee/tasting-notes
-tags:
-- '#coffee'
-- '#tasting'
-- '#flavor'
-- '#cupping'
-- '#demo'
----
-
-# Tasting Notes
-
-How to taste and evaluate coffee, identify flavor characteristics, and develop a personal coffee palate.
-
-## Overview
-
-Coffee tasting, or "cupping" in professional contexts, is the practice of observing the tastes and aromas of brewed coffee. Developing a coffee palate helps identify preferences, communicate about coffee experiences, and better understand how brewing variables affect the cup.
-
-## Observations
-
-- [principle] Flavor perception includes taste, aroma, mouthfeel, and retronasal perception #sensory
-- [principle] Our taste buds can only perceive sweet, sour, salty, bitter, and umami #taste
-- [principle] Most of what we call "flavor" is actually aroma detected retronasally #aroma
-- [technique] Professional coffee tasting (cupping) uses a standardized protocol for consistency #cupping
-- [technique] Slurping coffee aerates it and spreads it across all taste receptors #technique
-- [technique] Allowing coffee to cool reveals different flavor notes at different temperatures #temperature
-
-## Coffee Flavor Wheel
-
-- [tool] The SCA Coffee Flavor Wheel provides a standardized vocabulary for describing coffee #flavor_wheel
-- [category] Primary categories include: Fruity, Floral, Sweet, Nutty/Cocoa, Spice, Roasted, Other #categories
-- [subcategory] Fruity breaks down into: Berry, Dried Fruit, Citrus Fruit, Stone Fruit, Tropical Fruit, etc. #fruit_notes
-- [subcategory] Floral includes: Floral, Black Tea, Chamomile, Rose, Jasmine, etc. #floral_notes
-- [subcategory] Sweet includes: Brown Sugar, Molasses, Honey, Maple Syrup, Vanilla, etc. #sweet_notes
-- [subcategory] Nutty/Cocoa includes: Nut, Cocoa, Dark Chocolate, Chocolate, etc. #nutty_notes
-- [subcategory] Spice includes: Brown Spice, Pepper, Anise, Nutmeg, Cinnamon, etc. #spice_notes
-
-## Basic Tasting Components
-
-- [component] Acidity: The bright, tangy quality (not sourness from under-extraction) #acidity
-- [component] Sweetness: The pleasant, sugary quality balancing other elements #sweetness
-- [component] Body: The physical mouthfeel and weight of the coffee #body
-- [component] Finish/Aftertaste: The flavor that lingers after swallowing #finish
-- [component] Balance: How well all elements work together #balance
-- [component] Complexity: The range and layers of distinct flavors #complexity
-- [component] Cleanliness: Absence of defects or off-flavors #cleanliness
-
-## Common Flavor Notes by Origin
-
-- [ethiopia] Blueberry, jasmine, bergamot, lemon, tea-like #flavor_notes
-- [kenya] Blackcurrant, grapefruit, tomato-like acidity, winey #flavor_notes
-- [colombia] Caramel, red apple, nuts, chocolate, balanced acidity #flavor_notes
-- [guatemala] Chocolate, spice, apple, medium acidity #flavor_notes
-- [brazil] Nuts, chocolate, low-to-medium acidity, full body #flavor_notes
-- [indonesia] Earthy, herbal, spice, cedar, full body, low acidity #flavor_notes
-- [costa_rica] Clean, bright, citrus, balanced, light chocolate #flavor_notes
-
-## Developing Your Palate
-
-- [technique] Taste coffees side-by-side to identify differences #comparison
-- [technique] Try describing flavors before looking at roaster's notes #blind_tasting
-- [technique] Keep a coffee journal with detailed notes about each coffee #journaling
-- [technique] Explore different processing methods of the same origin #processing
-- [technique] Try the same coffee brewed with different methods #brewing_comparison
-- [technique] Use reference flavors (actual fruits, chocolates, etc.) to calibrate your palate #calibration
-
-## Personal Coffee Experiences
-
-- [experience] Ethiopian Yirgacheffe prepared as pour over: intense blueberry, jasmine aromatics, tea-like body
-- [experience] Sumatra Mandheling in French press: earthy, cedar, herbal, tobacco, full body
-- [experience] Panama Gesha as pour over: intense floral notes, jasmine, bergamot, delicate body
-- [experience] Brazil Cerrado as espresso: nutty, chocolate, caramel, low acidity, great crema
-- [experience] Kenya AA as pour over: bright blackcurrant, tomato-like acidity, winey finish
-
-## Relations
-
-- determined_by [[Flavor Extraction]]
-- influenced_by [[Coffee Bean Origins]]
-- varies_with [[Coffee Brewing Methods]]
-- enhanced_by [[Proper Grinding Technique]]
-- documented_in [[Coffee Journal]]
-- part_of [[Coffee Knowledge Base]]

examples/testing/Test Note Creation - Basic Functionality.md

@@ -1,69 +0,0 @@
----
-title: Test Note Creation - Basic Functionality
-type: note
-permalink: testing/test-note-creation-basic-functionality
-tags:
-- '["testing"'
-- '"core-functionality"'
-- '"note-creation"]'
----
-
----
-title: Test Note Creation - Basic Functionality
-tags: [testing, core-functionality, note-creation, edited]
-test_status: active
-last_edited: 2025-06-01
----
-
-# Test Note Creation - Basic Functionality
-
-## Test Status: COMPREHENSIVE TESTING IN PROGRESS
-Testing basic note creation with various content types and structures.
-
-## Content Types Tested
-- Plain text content ✓
-- Markdown formatting **bold**, *italic*
-- Lists:
-  - Bullet points
-  - Numbered items
-- Code blocks: `inline code`
-
-```python
-# Block code
-def test_function():
-    return "Hello, Basic Memory!"
-```
-
-## Special Characters
-- Unicode: café, naïve, résumé
-- Emojis: 🚀 🔬 📝
-- Symbols: @#$%^&*()
-
-## Frontmatter Testing
-This note should have proper frontmatter parsing.
-
-## Relations to Test
-- connects_to [[Another Test Note]]
-- validates [[Core Functionality Tests]]
-
-## Observations
-- [success] Note creation initiated
-- [test] Content variety included
-- [validation] Special characters included
-
-
-## Edit Test Results
-- [success] Note reading via title lookup ✓
-- [success] Search functionality returns relevant results ✓
-- [success] Special characters (unicode, emojis) preserved ✓
-- [test] Now testing append edit operation ✓
-
-## Performance Notes
-- Note creation: Instantaneous
-- Note reading: Fast response
-- Search: Good relevance scoring
-
-## Next Tests
-- Edit operations (append, prepend, find_replace)
-- Move operations
-- Cross-project functionality

justfile

@@ -2,25 +2,159 @@
 
 # Install dependencies
 install:
-    pip install -e ".[dev]"
+    uv pip install -e ".[dev]"
+    uv sync
+    @echo ""
+    @echo "💡 Remember to activate the virtual environment by running: source .venv/bin/activate"
 
-# Run unit tests in parallel
-test-unit:
-    uv run pytest -p pytest_mock -v -n auto
+# ==============================================================================
+# 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 integration tests in parallel
-test-int:
-    uv run pytest -p pytest_mock -v --no-cov -n auto test-int
+# Run all tests against SQLite and Postgres
+test: test-sqlite test-postgres
 
-# Run all tests
-test: test-unit 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
+test-int-sqlite:
+    uv run pytest -p pytest_mock -v --no-cov 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_TEST_POSTGRES=1 uv run pytest -p pytest_mock -v --no-cov test-int' || test $? -eq 137
+    else
+        echo "⚠️  No timeout command found, running without timeout..."
+        BASIC_MEMORY_TEST_POSTGRES=1 uv run pytest -p pytest_mock -v --no-cov 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:
+    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:
+    uv run pytest -p pytest_mock -v --no-cov -m benchmark tests test-int
+
+# 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:
+    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
 
 # Lint and fix code
-lint:
-    ruff check . --fix
+fix:
+    uv run ruff check --fix --unsafe-fixes src tests test-int
 
 # Type check code
-type-check:
+typecheck:
     uv run pyright
 
 # Clean build artifacts and cache files
@@ -38,26 +172,172 @@ format:
 run-inspector:
     npx @modelcontextprotocol/inspector
 
-# Build macOS installer
-installer-mac:
-    cd installer && chmod +x make_icons.sh && ./make_icons.sh
-    cd installer && uv run python setup.py bdist_mac
+# 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 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 type-check test
+check: lint format typecheck test
 
 # Generate Alembic migration with descriptive message
 migration message:
     cd src/basic_memory/alembic && alembic revision --autogenerate -m "{{message}}"
 
+# Create a stable release (e.g., just release v0.13.2)
+release version:
+    #!/usr/bin/env bash
+    set -euo pipefail
+    
+    # Validate version format
+    if [[ ! "{{version}}" =~ ^v[0-9]+\.[0-9]+\.[0-9]+$ ]]; then
+        echo "❌ Invalid version format. Use: v0.13.2"
+        exit 1
+    fi
+    
+    # Extract version number without 'v' prefix
+    VERSION_NUM=$(echo "{{version}}" | sed 's/^v//')
+    
+    echo "🚀 Creating stable release {{version}}"
+    
+    # Pre-flight checks
+    echo "📋 Running pre-flight checks..."
+    if [[ -n $(git status --porcelain) ]]; then
+        echo "❌ Uncommitted changes found. Please commit or stash them first."
+        exit 1
+    fi
+    
+    if [[ $(git branch --show-current) != "main" ]]; then
+        echo "❌ Not on main branch. Switch to main first."
+        exit 1
+    fi
+    
+    # Check if tag already exists
+    if git tag -l "{{version}}" | grep -q "{{version}}"; then
+        echo "❌ Tag {{version}} already exists"
+        exit 1
+    fi
+    
+    # Run quality checks
+    echo "🔍 Running lint  checks..."
+    just lint
+    just typecheck
+    
+    # Update version in __init__.py
+    echo "📝 Updating version in __init__.py..."
+    sed -i.bak "s/__version__ = \".*\"/__version__ = \"$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 commit -m "chore: update version to $VERSION_NUM for {{version}} release"
+    
+    # Create and push tag
+    echo "🏷️  Creating tag {{version}}..."
+    git tag "{{version}}"
+    
+    echo "📤 Pushing to GitHub..."
+    git push origin main
+    git push origin "{{version}}"
+    
+    echo "✅ Release {{version}} created successfully!"
+    echo "📦 GitHub Actions will build and publish to PyPI"
+    echo "🔗 Monitor at: https://github.com/basicmachines-co/basic-memory/actions"
+    echo ""
+    echo "📝 REMINDER: Post-release tasks:"
+    echo "   1. docs.basicmemory.com - Add release notes to src/pages/latest-releases.mdx"
+    echo "   2. basicmachines.co - Update version in src/components/sections/hero.tsx"
+    echo "   3. MCP Registry - Run: mcp-publisher publish"
+    echo "   See: .claude/commands/release/release.md for detailed instructions"
+
+# Create a beta release (e.g., just beta v0.13.2b1)
+beta version:
+    #!/usr/bin/env bash
+    set -euo pipefail
+    
+    # Validate version format (allow beta/rc suffixes)
+    if [[ ! "{{version}}" =~ ^v[0-9]+\.[0-9]+\.[0-9]+(b[0-9]+|rc[0-9]+)$ ]]; then
+        echo "❌ Invalid beta version format. Use: v0.13.2b1 or v0.13.2rc1"
+        exit 1
+    fi
+    
+    # Extract version number without 'v' prefix
+    VERSION_NUM=$(echo "{{version}}" | sed 's/^v//')
+    
+    echo "🧪 Creating beta release {{version}}"
+    
+    # Pre-flight checks
+    echo "📋 Running pre-flight checks..."
+    if [[ -n $(git status --porcelain) ]]; then
+        echo "❌ Uncommitted changes found. Please commit or stash them first."
+        exit 1
+    fi
+    
+    if [[ $(git branch --show-current) != "main" ]]; then
+        echo "❌ Not on main branch. Switch to main first."
+        exit 1
+    fi
+    
+    # Check if tag already exists
+    if git tag -l "{{version}}" | grep -q "{{version}}"; then
+        echo "❌ Tag {{version}} already exists"
+        exit 1
+    fi
+    
+    # Run quality checks
+    echo "🔍 Running lint  checks..."
+    just lint
+    just typecheck
+    
+    # Update version in __init__.py
+    echo "📝 Updating version in __init__.py..."
+    sed -i.bak "s/__version__ = \".*\"/__version__ = \"$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 commit -m "chore: update version to $VERSION_NUM for {{version}} beta release"
+    
+    # Create and push tag
+    echo "🏷️  Creating tag {{version}}..."
+    git tag "{{version}}"
+    
+    echo "📤 Pushing to GitHub..."
+    git push origin main
+    git push origin "{{version}}"
+    
+    echo "✅ Beta release {{version}} created successfully!"
+    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"
+
 # List all available recipes
 default:
-    @just --list
+    @just --list

memory.json

@@ -1,378 +0,0 @@
-{"type":"entity","name":"Paul","entityType":"person","observations":["Software developer combining DIY ethics, Free Software principles, and theoretical computer science","Created the Basic Machines project","Values authentic exchange of ideas","Approaches AI interaction with emphasis on genuine technical discussion","Comfortable with uncertainty and open dialogue","Balances practical implementation with broader implications"]}
-{"type":"entity","name":"Basic_Machines","entityType":"project","observations":["Local-first knowledge management system","Combines filesystem durability with graph-based knowledge representation","Focuses on enhancing human agency and understanding","Synthesizes DIY ethics, Free Software philosophy, and theoretical computer science","Current focus includes basic-memory system"]}
-{"type":"entity","name":"basic-memory","entityType":"software_system","observations":["A core component of Basic Machines","Local-first knowledge management system","Combines filesystem persistence with graph-based knowledge representation","Being implemented collaboratively by Paul and Claude"]}
-{"type":"entity","name":"basic-memory_implementation_patterns","entityType":"technical_patterns","observations":["Filesystem is source of truth - all changes write to files first","Clean separation of concerns between models (SQLAlchemy), schemas (Pydantic), and services","Repository pattern for database access","Service layer handling business logic and coordination","Atomic file operations using temporary files for safety","Clear error handling hierarchy with specific error types","Comprehensive test coverage with pytest and fixtures","Async/await used throughout the codebase","Validation using Pydantic models with custom validators"]}
-{"type":"entity","name":"fileio_module","entityType":"code_module","observations":["Extracted from EntityService to handle all file operations","Provides read_entity_file, write_entity_file, and delete_entity_file functions","Handles markdown parsing and formatting","Implements atomic file operations","Provides consistent error handling","Enables reuse across services"]}
-{"type":"entity","name":"entity_service","entityType":"code_module","observations":["Manages entities in both filesystem and database","Uses fileio module for file operations","Maintains database index of entities","Handles entity creation, retrieval, and deletion","Follows 'filesystem is source of truth' principle","Coordinates with observation service for full entity management"]}
-{"type":"entity","name":"observation_service","entityType":"code_module","observations":["Manages observations within entity files","Provides database indexing for efficient observation queries","Works with complete Entity objects rather than IDs","Handles observation addition and search","Maintains consistency between files and database","Under development for update/remove operations"]}
-{"type":"entity","name":"observation_management","entityType":"design_challenge","observations":["Key challenge: maintaining observation state across files and database","Exploring bulk update approach - treating all observations as a unit","Considering tracked observations with markdown comments for IDs","Investigating diff-based approach for observation-level changes","Evaluating position-based management without explicit IDs","Trade-offs between implementation complexity and markdown readability"]}
-{"type":"entity","name":"testing_infrastructure","entityType":"technical_patterns","observations":["Uses pytest with async support via pytest-asyncio","In-memory SQLite database for test isolation","Temporary directories for file operation testing","Comprehensive fixture system for test setup","Tests organized by component (entity, observation, etc)","Covers happy path, error cases, and edge cases","Uses monkeypatch for mocking dependencies","Clear separation between arrange, act, assert sections","Uses in-memory SQLite database for test isolation","Comprehensive fixture system for test data setup","Proper async test handling with pytest-asyncio"]}
-{"type":"entity","name":"test_categories","entityType":"test_suite","observations":["Happy path tests verify core functionality","Error path tests ensure proper error handling","Edge cases test special characters and long content","File operation tests verify atomic writes and rollbacks","Database sync tests verify index consistency","Recovery tests for rebuild operations","Punted on concurrent operation tests due to session management complexity"]}
-{"type":"entity","name":"completed_work","entityType":"project_milestone","observations":["Extracted file operations to fileio.py module","Updated EntityService to use fileio functions","Implemented initial ObservationService","Created comprehensive test suite","Established clear project patterns and principles","Set up basic database schema with SQLAlchemy","Created Pydantic models for validation"]}
-{"type":"entity","name":"future_work","entityType":"project_tasks","observations":["Implement observation updates/removals","Design proper session management for concurrent operations","Update EntityService tests for new fileio module","Add more sophisticated search functionality","Handle markdown formatting edge cases","Consider versioning for file changes","Implement proper backup strategy"]}
-{"type":"entity","name":"design_decisions","entityType":"technical_decisions","observations":["Filesystem as source of truth over database","Markdown format for human readability and editing","Atomic file operations for safety","SQLite + SQLAlchemy for proven reliability","Pydantic for validation and ID generation","Async/await for better scalability","Clear separation between files and database roles","Explicit error hierarchies for better handling"]}
-{"type":"entity","name":"concurrency_considerations","entityType":"technical_challenge","observations":["SQLAlchemy session management in async context","File operation atomicity","Transaction isolation levels","Potential for conflicting updates","Need for proper session lifecycle","Possibility of file system race conditions","Database lock management"]}
-{"type":"entity","name":"observation_update_approaches","entityType":"design_alternatives","observations":["Each approach trades off between simplicity, efficiency, and robustness","Four main approaches considered: bulk update, tracked IDs, diff-based, and position-based","Discussion revealed importance of human readability in file format","Consideration of manual editing workflows key to design","File system as source of truth principle guides tradeoffs"]}
-{"type":"entity","name":"bulk_update_approach","entityType":"design_option","observations":["Update all observations at once in a single operation","Simpler file operations - just rewrite the whole list","No need for observation matching or IDs","Very consistent with source of truth principle","Less efficient for small changes","May have concurrency implications","Simplest implementation option"]}
-{"type":"entity","name":"tracked_observations_approach","entityType":"design_option","observations":["Use markdown comments to store observation IDs","Enables precise updates and deletes","IDs stored as HTML comments in markdown","More complex markdown parsing required","IDs visible in raw markdown files","Balances tracking with readability"]}
-{"type":"entity","name":"diff_based_approach","entityType":"design_option","observations":["Implement observation-aware diffing","Track changes at observation level","More efficient for updates","Preserves manual edits and changes","More complex implementation needed","Must handle merge conflicts","Most sophisticated option considered"]}
-{"type":"entity","name":"position_based_approach","entityType":"design_option","observations":["Track observations by position/order","No explicit IDs needed","Cleanest markdown format","Order changes could break references","Difficult to handle concurrent edits","Most fragile option considered"]}
-{"type":"entity","name":"tasks_and_progress","entityType":"project_tracking","observations":["Current focus on observation management implementation","Completed core file operations extraction","Completed EntityService updates","Completed initial ObservationService","Basic test coverage in place","Future work includes concurrent operations","Future work includes search improvements","Need to handle markdown edge cases"]}
-{"type":"entity","name":"error_handling_patterns","entityType":"technical_patterns","observations":["Custom exception hierarchy with ServiceError base","Specific error types (FileOperationError, DatabaseSyncError, etc)","Clear separation between file and database errors","Error propagation patterns established","Focus on actionable error messages","Error handling at appropriate levels"]}
-{"type":"entity","name":"data_models","entityType":"technical_implementation","observations":["SQLAlchemy models for database structure","Pydantic schemas for API/service layer","Entity model with UUID-based IDs","Observation model with entity relationships","UTCDateTime custom type for timestamps","Automatic ID generation in Pydantic models","Strict validation rules"]}
-{"type":"entity","name":"markdown_format","entityType":"file_format","observations":["Simple, human-readable format","Entity name as H1 header","Metadata in key-value format","Observations as bullet points","Atomic file operations for updates","Designed for manual editing","No hidden metadata in main content"]}
-{"type":"entity","name":"test_driven_development","entityType":"development_pattern","observations":["Tests revealed need for atomic file operations","Error cases drove error hierarchy design","Edge cases informed validation rules","Test fixtures shaped service interfaces","File operations extracted due to test patterns","Concurrent test issues revealed session management needs"]}
-{"type":"entity","name":"architecture_evolution","entityType":"design_process","observations":["Started with simple EntityService implementation","Circular dependency between Entity and Observation services revealed design flaw","Extracted file operations to separate module","Moved to passing Entity objects rather than IDs","Improved separation of concerns through iterations","File operations became reusable across services","Database became true 'index' rather than source of truth"]}
-{"type":"entity","name":"validation_patterns","entityType":"technical_patterns","observations":["Pydantic models provide schema validation","Automatic ID generation if not provided","Database constraints via SQLAlchemy","Runtime checks in services","Markdown format validation","Error handling for invalid states"]}
-{"type":"entity","name":"markdown_examples","entityType":"documentation","observations":["Example of basic entity:\n# Entity Name\ntype: entity_type\n\n## Observations\n- First observation\n- Second observation","Example with special characters:\n# Test & Entity!\ntype: test\n\n## Observations\n- Test & observation with @#$% special chars!","Format ensures human readability:\n# Basic Machines\ntype: project\n\n## Observations\n- Local-first knowledge management system\n- Combines filesystem durability with graph-based knowledge representation","Future consideration for observation IDs:\n# Entity Name\ntype: entity_type\n\n## Observations\n- <!-- obs-id: abc123 -->\n  This is an observation with ID"]}
-{"type":"entity","name":"markdown_parsing_rules","entityType":"technical_implementation","observations":["H1 header contains entity name","Metadata uses key: value format","Observations section marked by H2 header","Each observation is a markdown list item","Blank lines separate sections","Special characters allowed in content","No restrictions on observation content"]}
-{"type":"entity","name":"schema_definitions","entityType":"technical_documentation","observations":["SQLAlchemy Entity model:\nclass Entity(Base):\n    id: str (primary key)\n    name: str (unique)\n    entity_type: str\n    created_at: datetime\n    updated_at: datetime","SQLAlchemy Observation model:\nclass Observation(Base):\n    id: str (primary key)\n    entity_id: str (foreign key)\n    content: str\n    created_at: datetime\n    context: Optional[str]","Pydantic Entity schema:\nclass Entity(BaseModel):\n    id: str\n    name: str\n    entity_type: str\n    observations: List[Observation]"]}
-{"type":"entity","name":"test_evolution","entityType":"development_history","observations":["Started with basic Entity CRUD tests","Added filesystem verification to all tests","Developed concurrent operation tests (later removed)","Edge case tests drove better error handling","Test fixtures evolved to support both file and DB testing","Mocking patterns for file/DB operations","Special cases for long content and special characters"]}
-{"type":"entity","name":"implementation_challenges","entityType":"technical_issues","observations":["Initial circular dependency between services","SQLAlchemy session management in async context","Atomic file operations with proper error handling","Maintaining DB sync with filesystem changes","Handling long content in observations","Managing test isolation with file operations","Deciding on markdown format tradeoffs","Concurrent operation complexity"]}
-{"type":"entity","name":"Basic_Factory","entityType":"Project","observations":["Collaborative project between Paul and Claude","Explores AI-human collaboration in software development","Uses MCP tools for file and memory management","Built with git integration capabilities","Focuses on maintaining project context across sessions","About 90% complete with MCP tools","Still needs improvements in collaboration via files/git/github","Will be used to document and share collaborative development process"]}
-{"type":"entity","name":"Basic_Factory_Components","entityType":"Technical","observations":["Server-side rendering with JinjaX","HTMX for dynamic updates","Alpine.js for client-side state","Tailwind CSS for styling","Component translation from React/shadcn/ui","Focus on simplicity and understandability","Demonstrates meta-compiler principles in component translation"]}
-{"type":"entity","name":"Component_Translation_Process","entityType":"Methodology","observations":["Treats component porting as meta-compilation","Maps between React/TypeScript and JinjaX/Alpine.js domains","Uses formal grammar transformation approaches","Maintains functionality while simplifying implementation","Focuses on server-side rendering patterns","Preserves accessibility and performance","Uses short, focused git branches for each component"]}
-{"type":"entity","name":"Basic_Machines_Philosophy","entityType":"Philosophy","observations":["Combines DIY punk ethics with software development","Emphasizes user empowerment and understanding","Values simplicity and composability","Treats complex systems as combinations of simple parts","Focuses on authentic creation and sharing","Draws inspiration from punk rock, Free Software, and theoretical CS","Emphasizes the cycle of creation, complexity, and renewal"]}
-{"type":"entity","name":"Basic_Machines_Manifesto","entityType":"Document","observations":["Created through collaboration between Paul and Claude","Explores connection between DIY punk ethics and software development","Emphasizes composition over inheritance in both philosophy and practice","Views software development through lens of basic machines that combine for complex computation","Advocates for user empowerment and technological independence","Structured in sections covering Origins, Philosophy, Technical Implementation, and AI Collaboration","Draws connections between punk rock, free software, and theoretical computer science","Emphasizes importance of sharing knowledge and building community","Released in December 2024"]}
-{"type":"entity","name":"AI_Human_Collaboration_Model","entityType":"Methodology","observations":["Focuses on deep collaboration rather than simple task completion","Maintains rich context across sessions via knowledge graph","Uses short, focused git branches for each collaborative session","Values intellectual partnership over simple code generation","Emphasizes both practical implementation and theoretical exploration","Creates space for authentic exchange while maintaining AI/human clarity","Uses formal methods when appropriate (like grammar transformation)","Documents decisions and processes for future reference","Developed through Basic Machines project experience"]}
-{"type":"entity","name":"Basic_Machines_Roadmap","entityType":"Project_Plan","observations":["Phase 1 (30 days): Build basic-machines.co website","Phase 2 (60-90 days): Develop premium component bundles","Phase 3 (90-120 days): Launch Basic Foundation commercial offering","Focus on building brand and marketing presence","Prioritize components needed for own site development","Document and share collaboration process","Build sustainable business model aligned with values"]}
-{"type":"entity","name":"Basic_Machines_Website","entityType":"Project","observations":["To be built at basic-machines.co","Will showcase products and vision","Needs components for navigation, hero sections, features","Will demonstrate component usage in production","Will include blog for sharing progress","Focus on clear value proposition","Platform for sharing Basic Machines philosophy"]}
-{"type":"entity","name":"Basic_Memory_Markdown_Example","entityType":"Example","observations":["Shows complete markdown structure for basic-memory entity","Uses frontmatter for metadata (id, type, created, context)","Has main description section after title","Includes Observations as bullet points","Shows Relations with [id] relation_type | context format","Lists References at bottom","Created during initial design discussion","Serves as canonical example of file format"]}
-{"type":"entity","name":"Basic_Memory_Database_Schema","entityType":"Technical","observations":["Uses SQLite for local storage","Entities table with id, name, type, created_at, context, description, references","Observations table linking to entities with content and context","Relations table tracking directional relationships between entities","References column needs quotes as SQL reserved word","Designed for easy rebuilding from markdown files","Foreign key constraints maintain data integrity","Unique constraint on relations prevents duplicates","Created_at timestamps track history","Context fields enable tracking information sources"]}
-{"type":"entity","name":"Basic_Memory_Project_Structure","entityType":"Technical","observations":["Uses dbmate for database migrations","Projects directory stores SQLite databases and markdown files","Makefile provides common development commands","Environment vars configure database connection","db/migrations directory for SQL schema changes","Gitignore excludes database files and env config","Uses Python 3.12 with modern tooling","Tests directory for pytest files","Follows Basic Machines project conventions"]}
-{"type":"entity","name":"Basic_Memory_Project_Isolation_Decision","entityType":"Decision","observations":["Decided to defer multi-project support to post-MVP","Will use separate SQLite databases per project","Initially using projects directory in code repository","Plan to make location configurable later","No changes needed to core domain model","Keeps initial implementation simple","FTS/search capabilities also deferred for simplicity"]}
-{"type":"entity","name":"Basic_Memory_Implementation_Plan","entityType":"Plan","observations":["Start with SQLAlchemy models matching schema","Then build CLI for basic operations","Then implement markdown parser","Use TDD approach throughout","Begin with core domain model","CLI will support CRUD operations","Parser must handle frontmatter and sections","Following modular development approach","Planning to use typer for CLI","Will use modern Python tools and practices"]}
-{"type":"entity","name":"Basic_Memory_Implementation_Status","entityType":"Status","observations":["Core modules implemented: models, services, repository, fileio","Modular architecture with clear separation of concerns","File operations extracted to separate fileio module","Initial ObservationService implementation complete","Basic test coverage in place","Exploring observation management strategies","Using SQLAlchemy for database interaction","Markdown file operations working","Entity management functional","Repository layer implementation complete with SQLAlchemy models and tests","Database operations working with proper UTC timestamp handling","In-memory SQLite testing infrastructure proven effective"]}
-{"type":"entity","name":"Basic_Memory_Observation_Management_Design","entityType":"Design","observations":["Four approaches under consideration","Bulk Update: Simple but less efficient","Tracked Observations: Precise but clutters markdown","Diff-based: Efficient but complex","Position-based: Clean but fragile","Key challenge is balancing markdown readability with efficient updates","Must maintain filesystem as source of truth","Need to consider concurrent edits","Currently evaluating trade-offs","Implementation choice pending discussion"]}
-{"type":"entity","name":"Basic_Memory_Architectural_Decisions","entityType":"Decisions","observations":["Split file operations into separate fileio module","Using SQLAlchemy for database operations","Maintain filesystem as source of truth","Modular service-based architecture","Clear separation between data access and business logic","Repository pattern for database interactions","Schemas separate from models","Focus on maintainability and testability","Services handle business rules","Considering concurrency in design"]}
-{"type":"entity","name":"Basic_Memory_Implementation_Analysis","entityType":"Analysis","observations":["Clean modular architecture with clear responsibilities","Strong typing throughout codebase","Excellent error handling with custom exceptions","SQLAlchemy models perfectly match our domain model","Atomic file operations for data safety","Services implement filesystem-as-source-of-truth principle","Async support throughout","Good separation between domain models and database models","Careful handling of UTC timestamps","Smart use of SQLAlchemy relationships"]}
-{"type":"entity","name":"Basic_Memory_Current_Challenges","entityType":"Challenges","observations":["Observation update/removal strategy needs to be chosen","Need to handle concurrent file operations safely","Search functionality to be implemented","Edge cases in markdown formatting to be handled","Session management for concurrent operations needed","Balance between file operations and database sync","Testing coverage could be expanded","Need to handle relationship updates in files"]}
-{"type":"entity","name":"Basic_Memory_Observation_Hash_Tracking","entityType":"Design","observations":["Use content hashes to track observation identity","Store hashes in database but not in markdown","Can match observations across file edits using hashes","Similar to how git tracks content changes","Keeps markdown clean and human-friendly","Allows efficient bulk updates","Handles reordering of observations","Maintains filesystem as source of truth","No need for visible IDs in markdown","Could track observation history through hash changes"]}
-{"type":"entity","name":"Basic_Memory_Repository_Implementation","entityType":"Code_Implementation","observations":["Implemented base Repository class with CRUD operations","Added specialized EntityRepository, ObservationRepository, and RelationRepository","Used string IDs instead of UUIDs","Added UTCDateTime custom type for timestamp handling","Used in-memory SQLite for testing","Achieved 84% test coverage","Created comprehensive pytest fixtures"]}
-{"type":"entity","name":"Basic_Memory_Dependencies","entityType":"Technical","observations":["Uses Python 3.12","SQLAlchemy with async support","pytest-asyncio for async testing","aiosqlite for async SQLite operations","greenlet for SQLAlchemy async support","uv for dependency management","pytest-cov for coverage reporting","Development dependencies managed in pyproject.toml"]}
-{"type":"entity","name":"Basic_Memory_Current_Architecture","entityType":"Architecture_Analysis","observations":["Clear separation between domain models (Pydantic) and storage models (SQLAlchemy)","File I/O completely separated into dedicated module","Strong 'filesystem as source of truth' pattern in services","Atomic file operations with proper error handling","Service layer coordinates between filesystem and database","Database acts as queryable index rather than primary storage","Clean error hierarchy with specific exception types","Rebuild operations available for recovery scenarios"]}
-{"type":"entity","name":"Basic_Memory_Evolution","entityType":"Analysis","observations":["Started with repository pattern following basic-foundation","Evolved to more sophisticated architecture with clear layers","Added Pydantic schemas for domain modeling","Separated file operations into dedicated module","Implemented robust error handling throughout","Maintained filesystem as source of truth principle","Added observation management with context tracking","Introduced rebuild capabilities for system recovery"]}
-{"type":"entity","name":"Basic_Memory_Service_Layer","entityType":"Implementation","observations":["EntityService handles entity lifecycle and coordinates storage","ObservationService manages observations within entities","Services ensure filesystem and database stay in sync","Clear error handling with ServiceError hierarchy","Strong typing throughout service interfaces","Implements filesystem as source of truth pattern","Handles UUID generation and timestamp management","Provides methods for system recovery and rebuild"]}
-{"type":"entity","name":"Basic_Memory_Schema_Design","entityType":"Implementation","observations":["Uses Pydantic for domain models and validation","Automatic ID generation with timestamp and UUID","Clear separation from SQLAlchemy storage models","Supports optional context tracking","Models match markdown file structure","Enables clean serialization/deserialization","Strong typing with proper validation rules","Independent from storage concerns"]}
-{"type":"entity","name":"Basic_Memory_Next_Tasks","entityType":"TaskList","observations":["✅ Implement SQLAlchemy models and repositories (Done)","✅ Add SQLAlchemy migrations (Done)","✅ Create service layer (Done)","✅ Implement file I/O module (Done)","✅ Set up domain models with Pydantic (Done)","✅ Initial test infrastructure (Done)","✅ Basic CRUD operations (Done)","⏳ Implement full test coverage for db.py","⏳ Add more sophisticated search functionality","⏳ Implement CLI interface","⏳ Add relationship management to services","⏳ Handle concurrent file operations safely","⏳ Add versioning for file changes","⏳ Implement proper backup strategy","⏳ Add type hints throughout codebase","⏳ Improve error messages and logging","⏳ Add documentation for core modules"]}
-{"type":"entity","name":"Basic_Memory_Meta_Experience","entityType":"Case_Study","observations":["Experienced our own context loss when reconstructing project knowledge","Had to rebuild task list and project context from filesystem and memory","Validated 'filesystem as source of truth' principle through reconstruction","Code and tests served as reliable historical record","Knowledge graph structure helped guide reconstruction process","Markdown files provided human-readable context","Atomic information design made piece-by-piece reconstruction possible","Ironic validation of the need for basic-memory's features","Experience demonstrates value of durable, human-readable knowledge storage","Shows importance of separating durable storage from ephemeral context"]}
-{"type":"entity","name":"Model_Context_Protocol","entityType":"protocol","observations":["Core part of the basic-memory architecture","Enables AI-human collaboration on projects","Provides tool-based interaction with knowledge graph","Developed by Anthropic for structured AI-system interaction","Used for maintaining consistent, rich context across conversations"]}
-{"type":"entity","name":"basic-memory_core_principles","entityType":"principles","observations":["Local First: All data stored locally in SQLite","Project Isolation: Separate databases per project","Human Readable: Everything exportable to plain text","AI Friendly: Structure optimized for LLM interaction","DIY Ethics: User owns and controls their data","Simple Core: Start simple, expand based on needs","Tool Integration: MCP-based interaction model"]}
-{"type":"entity","name":"basic-memory_business_model","entityType":"business_strategy","observations":["Core features free: Local SQLite, basic knowledge graph, search, markdown export, basic MCP tools","Professional features potential: Rich document export, advanced versioning, collaboration features, custom integrations, priority support","Focus on maintaining DIY/punk philosophy while enabling sustainability"]}
-{"type":"entity","name":"basic-memory_cli","entityType":"interface","observations":["Supports project management commands (create, switch, list)","Entity management (add entity, add observation, add relation)","Future support for export and batch operations","Follows consistent command structure","Planned integration with MCP tools"]}
-{"type":"entity","name":"basic-memory_export_format","entityType":"file_format","observations":["Uses markdown with frontmatter metadata","Includes entity name, type, creation timestamp","Observations as bullet points","Relations in structured format with links","References section at bottom","Designed for human readability and machine parsing","Example format documented in project specs"]}
-{"type":"entity","name":"relation_service","entityType":"code_module","observations":["Planned service for managing relations in both filesystem and database","Will follow filesystem-is-source-of-truth principle like other services","Needs to handle atomic file operations for relation updates","Must coordinate with EntityService for relationship integrity","Will handle bidirectional relationship tracking","Will support relation validation and type enforcement","Must implement rebuild functionality for index recovery","Will need careful error handling for file/db sync","Should support relation search and filtering","Must handle relation lifecycle (create/read/update/delete)"]}
-{"type":"entity","name":"service_layer_patterns","entityType":"implementation_patterns","observations":["Services handle both file and database operations","Filesystem is always source of truth","Database serves as queryable index","Services implement atomic file operations","Clear error hierarchy with specific exceptions","Use of dependency injection via constructor params","Async/await used throughout service layer","Services coordinate between storage layers","Repository pattern used for database access","Services maintain entity integrity across storage","Rich error types extend from ServiceError base","Rebuild operations available for recovery"]}
-{"type":"entity","name":"database_models","entityType":"implementation","observations":["Entity model with unique name and type","Observation model linked to entities","Relation model tracks connections between entities","Custom UTCDateTime type for timestamp handling","Use of SQLAlchemy relationships for navigation","Cascading deletes for dependent objects","String IDs used for compatibility","Rich relationship modeling with backpopulates","Proper indexing on foreign keys","Context tracking available on models","Models include created_at timestamps","Relationships handle bidirectional navigation"]}
-{"type":"entity","name":"repository_patterns","entityType":"implementation_patterns","observations":["Generic Repository[T] base class implementation","Type-safe operations with SQLAlchemy","Specialized repositories for each model type","Async operations throughout","Clear error handling patterns","Support for custom queries and filtering","Pagination support built-in","Transaction management via session","Proper type hints and generics usage","Entity-specific query methods in subclasses"]}
-{"type":"entity","name":"relation_service_design","entityType":"design","observations":["Must handle relation lifecycle in both files and DB","Needs to validate existence of both entities","Should support relation type enforcement","Must maintain bidirectional consistency","Should support relation querying and filtering","Needs proper error handling for graph consistency","Must integrate with entity file format","Should support bulk operations for efficiency","Must handle relation deletion and cascading","Should provide search by type and entities"]}
-{"type":"entity","name":"relation_service_implementation_plan","entityType":"plan","observations":["1. Define core relation operations (create, get, delete)","2. Implement file format handling for relations","3. Add database sync with RelationRepository","4. Implement validation and error handling","5. Add rebuild and recovery operations","6. Implement relation type enforcement","7. Add relation search and filtering","8. Implement bulk operations","9. Add comprehensive tests","10. Document API and error handling"]}
-{"type":"entity","name":"relation_service_challenges","entityType":"challenges","observations":["Maintaining consistency between file and database","Handling relation type validation efficiently","Managing bidirectional relationships in files","Ensuring atomic updates across entities","Handling deletion with proper cascading","Efficient querying of relation graphs","Recovery from partial file/db sync failures","Bulk operation atomicity","Clear error reporting for graph operations","Performance with large relation sets"]}
-{"type":"entity","name":"relation_file_format","entityType":"file_format","observations":["Relations stored in entity markdown files","Format: [target_id] relation_type | context","Relations section marked by ## Relations header","Outgoing relations only stored in source entity","Relations rebuild on entity load","Clean human-readable format","Context is optional with pipe separator","Links generate valid navigation references","Markdown-friendly formatting","Example: [Paul] authored | with Claude"]}
-{"type":"entity","name":"relation_service_error_handling","entityType":"implementation_patterns","observations":["RelationError extends ServiceError base","Specific errors for validation failures","Handles entity not found cases","Manages relation type validation errors","File operation errors properly wrapped","Database sync errors clearly reported","Transaction rollback on errors","Proper error propagation chain","Clear error messages for debugging","Recovery paths for common errors"]}
-{"type":"entity","name":"relation_service_testing","entityType":"testing","observations":["Test all relation lifecycle operations","Verify file and database consistency","Test relation type validation","Check error handling paths","Test bulk operations","Verify bidirectional consistency","Test recovery operations","Check cascade operations","Verify search and filtering","Test with large relation sets"]}
-{"type":"entity","name":"fileio_patterns","entityType":"implementation_patterns","observations":["Atomic file operations with temporary files","Clear error handling for IO operations","Consistent file naming and paths","Support for different file formats","Efficient file reading and writing","Proper file locking mechanisms","Recovery from partial writes","Consistent encoding handling","Directory management utilities","Path manipulation helpers","Currently implemented in fileio.py module","Uses pathlib for path operations","Handles file not found cases gracefully","Maintains data integrity during writes"]}
-{"type":"entity","name":"pytest_patterns","entityType":"implementation_patterns","observations":["Common fixtures should be in conftest.py for reuse","Use pytest_asyncio.fixture for async fixtures","Session fixtures need proper async cleanup","Temporary directories should be managed with context managers","Test categories: happy path, error path, recovery, edge cases","Services need project_path and repo injected","Use monkeypatch for mocking in async context","SQLite in-memory database ideal for testing","Explicit test verification: file content and database state"]}
-{"type":"entity","name":"relation_implementation_learnings","entityType":"implementation_learnings","observations":["Better to pass full Entity objects than IDs to services","Services should not re-read entities if they have them","File operations should be atomic and verified","Database serves as queryable index, not source of truth","Relations stored in source entity's markdown file","Clear separation between file ops and database sync","Entity objects should own their relations list","Context is optional but fully supported in implementation"]}
-{"type":"entity","name":"test_driven_insights","entityType":"learnings","observations":["Tests help reveal better API design (e.g., passing Entity objects)","Error cases drive proper exception hierarchy","File verification as important as database checks","Edge cases inform markdown format decisions","Recovery tests ensure system resilience","Tests document expected behavior clearly","Fixtures significantly reduce test complexity","Common patterns emerge through test writing"]}
-{"type":"entity","name":"meta_development_insights","entityType":"process","observations":["Break down large tasks into reviewable chunks","One file at a time prevents response truncation","Iterative development with tests leads to better design","Infrastructure code (fixtures) should be consolidated early","Test categories help ensure comprehensive coverage","Knowledge capture should happen during development","APIs tend to evolve toward simpler patterns","File operations require careful verification"]}
-{"type":"entity","name":"AI_Assistant_Learnings","entityType":"meta_insights","observations":["Output management: Breaking responses into single files prevents truncation and allows better review","Knowledge graph helps maintain context: I can reference previous decisions and patterns accurately","Memory rebuilding experience validated the need for durable storage","Test-driven development provides clear steps and verification","Explicit relation tracking in knowledge graph helps me understand project context","Rich context from multiple sources (code, docs, tests) enables better assistance","File-at-a-time approach allows deeper analysis of each component","Keeping entity names consistent helps with referencing and relationships"]}
-{"type":"entity","name":"Effective_Response_Patterns","entityType":"meta_patterns","observations":["When showing code changes, break into discrete files","Review existing code before suggesting changes","Reference knowledge graph for context and patterns","Explicitly connect new code to existing patterns","Validate suggestions against test cases","Keep track of file changes for atomic commits","Check both implementation and test files for consistency","Maintain clear separation of concerns in responses"]}
-{"type":"entity","name":"AI_Context_Management","entityType":"meta_practice","observations":["Knowledge graph provides reliable persistent memory","Project documentation gives high-level context","Code review shows implementation patterns","Tests demonstrate expected behavior","Important to actively track what has been modified","Entity relationships help understand dependencies","Regular knowledge capture during development","Using consistent entity references across conversations"]}
-{"type":"entity","name":"AI_Tool_Usage_Patterns","entityType":"meta_practice","observations":["read_file before suggesting changes","write_file one file at a time","list_directory to understand project structure","search_nodes to find relevant context","create_entities to capture new learnings","create_relations to connect concepts","Using knowledge graph to track decisions","Validating changes through test execution"]}
-{"type":"entity","name":"relation_service_learnings","entityType":"implementation_learnings","observations":["Entity-based API cleaner than ID-based for service layer","Model_dump method can handle storage serialization","File format needs explicit section markers (## Relations)","Whitespace handling important for long content comparisons","Test fixtures allow focused test cases","SQLAlchemy selects better than raw SQL for type safety","Atomic file operations maintained for relations"]}
-{"type":"entity","name":"test_driven_insights_relations","entityType":"learnings","observations":["Tests revealed need for whitespace normalization","Edge cases drove file format decisions","SQLAlchemy model access safer than raw queries","Fixtures reduced test setup complexity","File verification as important as database checks","Testing both memory model and storage format","Test categories ensure comprehensive coverage"]}
-{"type":"entity","name":"relation_service_patterns","entityType":"patterns","observations":["Use Entity objects in API","Serialize to IDs for storage","Maintain file as source of truth","Keep file format human-readable","Handle circular references in serialization","Use repository pattern for database","Clear error hierarchies"]}
-{"type":"entity","name":"packaging_learnings","entityType":"technical_learnings","observations":["When using pytest-mock, traditional pip install works more reliably than uv sync","Package discovery behavior can differ between uv and pip","Clean venv with pip install is a reliable fallback for dependency issues","Package installation location might differ between uv and pip","Dependencies in pyproject.toml dev section work reliably with pip install -e .[dev]"]}
-{"type":"entity","name":"Recent_Implementation_Progress","entityType":"progress_update","observations":["Successfully split services.py into modular structure under services/","Created __init__.py, entity_service.py, observation_service.py, relation_service.py","Fixed pytest-mock installation issues by using pip install -e .[dev] instead of uv sync","Improved test structure with minimal mocking - only used for error testing","Implemented relation service with Entity-based API","Achieved good test coverage across services","File operations are only mocked when testing error conditions","Services follow filesystem-as-source-of-truth pattern"]}
-{"type":"entity","name":"Next_Steps","entityType":"project_tasks","observations":["Consider adding more relation service tests","Potentially expand relations features","Look for opportunities to improve test coverage","Consider documenting package management preferences (pip vs uv)","Consider adding integration tests for services","Review and possibly expand error handling cases"]}
-{"type":"entity","name":"Development_Practices","entityType":"process","observations":["Favor real operations over mocks in tests","Only mock for error condition testing","Use pip install -e .[dev] for reliable dev dependency installation","Maintain modular service structure","Keep filesystem as source of truth","Use Entity objects in service APIs instead of IDs","Validate both file and database state in tests"]}
-{"type":"entity","name":"MCP_Resources","entityType":"Concept","observations":["Stateful objects in Model Context Protocol","Enable persistent access to capabilities"]}
-{"type":"entity","name":"MCP_Server_Implementation","entityType":"Technical_Design","observations":["Inherits from mcp.server.Server base class","Tools are implemented as async methods","Each tool method maps directly to a function available to the AI","Tools can request user input via Prompts","Simple function call interface rather than explicit resource management","State management handled by server instance","Returns serialized data using model_dump() for consistency"]}
-{"type":"entity","name":"MCP_Tools","entityType":"Protocol_Feature","observations":["Defined as async methods on server class","Return values must match tool definition schema","Can maintain state between invocations via server instance","Tools can prompt for user input when needed","No need for explicit Resource objects in implementation"]}
-{"type":"entity","name":"Basic_Memory_MCP","entityType":"Implementation","observations":["Uses MemoryService for core operations","Implements project selection via prompts","Maintains project context across tool invocations","Maps directly to memory graph operations","Handles serialization of Pydantic models"]}
-{"type":"entity","name":"Basic_Memory_Testing","entityType":"Testing_Design","observations":["Needs pytest for async testing","Should isolate filesystem operations for tests","Needs to handle MCP server lifecycle in tests","Should test both service layer and MCP interface","Will need mocks for project paths and file operations"]}
-{"type":"entity","name":"Memory_Service_Tests","entityType":"Test_Suite","observations":["Should test entity creation with observations","Should test relation creation between entities","Should verify proper ID generation and model validation","Should test deletion cascading","Should test search functionality","Must verify proper serialization of entities and relations"]}
-{"type":"entity","name":"MCP_Server_Tests","entityType":"Test_Suite","observations":["Should test project initialization workflow","Should test prompt handling","Should verify tool input/output formats","Should test error cases and validation","Must verify proper serialization in tool responses"]}
-{"type":"entity","name":"Memory_Service_Refactoring","entityType":"Technical_Task","observations":["MemoryService uses create() but EntityService might expect create_entity()","MemoryService assumes get_by_name() but EntityService might use different method","Need to verify deletion method signatures","Need to check if search interface matches","Should verify observation handling matches ObservationService interface","RelationService methods need verification","EntityService.create_entity takes name, type, and optional observations directly, not an Entity object","EntityService requires project_path and entity_repo in constructor","ObservationService.add_observation takes Entity object and content string, not raw data","RelationService.create_relation takes Entity objects directly, not dict data","All services follow filesystem-as-source-of-truth pattern with DB indexing","All services handle database synchronization internally","Services expect Path objects for filesystem operations"]}
-{"type":"entity","name":"Service_Interface_Audit","entityType":"Technical_Task","observations":["Need to review all existing service interfaces","Document current method signatures","Map discrepancies between MemoryService assumptions and actual interfaces","Check return types and error handling patterns","Review transaction/atomicity requirements","Method signatures need alignment: create vs create_entity etc","Need to handle DB repositories in service constructors","File operations should use project_path consistently","Need to maintain filesystem-as-source-of-truth pattern","Should handle database synchronization at service level","Error handling should align with existing patterns","Consider making MemoryService handle DB indexing consistently"]}
-{"type":"entity","name":"Memory_Service_Patterns","entityType":"Technical_Pattern","observations":["Uses inner async functions to encapsulate operation logic","Leverages list comprehensions with async functions for parallel operations","Each operation follows a consistent pattern: validate, update DB, write file","Inner functions make the code more readable and maintainable","Operations can run in parallel when using list comprehensions with async functions"]}
-{"type":"entity","name":"Pydantic_Create_Pattern","entityType":"Technical_Pattern","observations":["Separate Create models match the exact shape of incoming data","Provides clear contract for MCP tool inputs","Handles validation of raw input data","Converts cleanly to domain models via from_create methods","Maintains separation between external API format and internal models","Similar to FastAPI request model pattern","Allows camelCase in API while using snake_case internally"]}
-{"type":"entity","name":"Basic_Memory_Business","entityType":"Business_Model","observations":["Core system is open source and free","Local-first, giving users data control","Professional features could be licensed","Enterprise support and customization services","Potential for MCP tool marketplace"]}
-{"type":"entity","name":"MCP_Marketplace","entityType":"Business_Concept","observations":["Could host verified MCP tools for different use cases","Tools rated by performance and reliability","Marketplace takes percentage of tool usage fees","Enterprise tool verification and security scanning","Custom tool development services","Integration support for existing tools"]}
-{"type":"entity","name":"Persistence_Of_Vision","entityType":"Concept","observations":["Mental model for continuous AI-human interaction","Like cinema: 24fps creates illusion of smooth motion","Basic-memory provides 'frames' of structured knowledge","Current state: Better than flipbook, not yet digital cinema","Goal: Achieve smoother cognitive continuity between interactions","Proposed by Drew as metaphor for AI conversation continuity"]}
-{"type":"entity","name":"Conversation_Continuity_Pattern","entityType":"Usage_Pattern","observations":["Use basic-memory entity/relation schema for conversations","Each chat becomes an entity with observations for key points","Relations link to discussed concepts and other chats","Uses zettelkasten format IDs for natural ordering","Can be used as template/recipe for others","Future possibility: Git SHA integration for versioning"]}
-{"type":"entity","name":"Usage_Recipes","entityType":"Feature_Concept","observations":["Predefined patterns users can follow or adapt","Could include conversation tracking recipe","Templates for different knowledge management styles","Shows practical applications of the generic schema","Helps users get started with the system"]}
-{"type":"entity","name":"Chat_References","entityType":"Technical_Feature","observations":["Uses ref:* syntax to reference previous conversations","Combines reference semantics with pointer symbolism","Format: ref:*{zettelkasten-id}","Allows explicit context loading between chats","Inspired by C++ references and pointers","Provides memory-model-like access to conversation context","Uses ref:// URI format following MCP Resource pattern","Could support multiple reference schemes (chat/entity/concept)","Makes reference semantics explicit and unambiguous","Aligns with standard URI formatting"]}
-{"type":"entity","name":"Chat_Reference_Protocol","entityType":"Technical_Specification","observations":["Uses URI format: ref://basic-memory/chat/[id]","Follows MCP Resource pattern: [protocol]://[host]/[path]","Enables explicit context loading between chats","Can support multiple resource types (chat/entity/concept)","Provides standardized way to reference previous conversations","Example: ref://basic-memory/chat/20240307-drew-ab12ef34"]}
-{"type":"entity","name":"20240307-chat-reference-protocol","entityType":"conversation","observations":["Developed ref:// URI format for chat references","Added Chat Reference Protocol to prompt instructions","Discussed implementation of chat continuation","Created complete prompt instructions document","Reference format follows MCP Resource pattern","Reviewed and confirmed complete prompt instructions","Ready to test ref://basic-memory/chat/20240307-chat-reference-protocol in new chat"]}
-{"type":"entity","name":"20240307-chat-reference-protocol-test","entityType":"conversation","observations":["First implementation test of chat reference protocol","Testing continuation from 20240307-chat-reference-protocol","Focused on practical implementation of ref:// URI format"]}
-{"type":"entity","name":"Write_File_Tool_Usage","entityType":"Tool_Usage_Pattern","observations":["Never use placeholders like '# Rest of...' when writing files - must include complete file content","File content must be complete and valid - partial updates will truncate the file","If showing partial changes, should inform human and let them handle the file write","write_file tool replaces entire file contents - cannot do partial updates","Code files especially must be complete and valid to avoid breaking functionality","Always read_file before write_file to understand current state","Using write_file without reading first risks reverting recent changes","Pattern should be: read current state, make modifications, then write if needed","Especially important in collaborative development where files may have been updated"]}
-{"type":"entity","name":"Run_Tests_Tool_Request","entityType":"Feature_Request","observations":["Need to add a tool enabling Claude to run tests locally","Would help with direct validation of code changes","Current workaround: Claude has to ask human to run tests","Should support running specific test functions (e.g. pytest tests/test_memory_service.py::test_create_relations)","Would improve iterative development workflow between human and AI"]}
-{"type":"entity","name":"SQLAlchemy_Async_Loading_Pattern","entityType":"Technical_Pattern","observations":["Use selectinload() instead of lazy loading when accessing SQLAlchemy relationships in async code","Lazy loading doesn't work with async due to greenlet context requirements","selectinload performs a single efficient query with an IN clause","Pattern used in basic-memory's EntityRepository for loading relations","Documented in find_by_id method with thorough explanation","Alternative approaches: joinedload (single JOIN query) or subqueryload (subquery approach)","Benefits: prevents 'MissingGreenlet' errors, reduces N+1 query problems","Key insight: load all needed relationships upfront in async code","Example use: selectinload(Entity.outgoing_relations)"]}
-{"type":"entity","name":"20241207-sqlalchemy-async-pattern","entityType":"conversation","observations":["Fixed SQLAlchemy async relationship loading issues","Implemented selectinload pattern in EntityRepository","Updated find_by_id to eager load relations","Added documentation about the pattern","Created knowledge graph entry about SQLAlchemy async loading","Fixed failing tests by properly loading relations in memory_service","Discussed SQLAlchemy relationship loading best practices"]}
-{"type":"entity","name":"20241207-memory-service-relations","entityType":"conversation","observations":["Fixed SQLAlchemy async loading with selectinload pattern","Updated find_by_id in EntityRepository to eager load relations","Discovered create_relations works but returns empty list","Verified relations are being stored correctly in memory.json","Next step: Work on MemoryService.add_observations implementation","Improved understanding of MCP memory storage format through debugging"]}
-{"type":"entity","name":"add_observations_implementation_plan","entityType":"technical_plan","observations":["Follow pattern from create_entity and create_relation methods","File operations first (read & write) - filesystem is source of truth","Database updates in parallel","Simplify current implementation","Current flow is:"," - First read entities and create observations"," - Write files in parallel"," - Update DB indexes sequentially","Key tests needed:"," - Adding observations to multiple entities"," - Verifying filesystem state first"," - Verifying database state"," - Error cases for missing entities"," - Error cases for file operations"]}
-{"type":"entity","name":"MCP_Reference_Integration","entityType":"feature_idea","observations":["Can be implemented as a Model Context Protocol integration similar to the fetch tool","Would provide structured way to pass chat references to Claude","Could handle ref:// URL format systematically","Integration would fetch context from referenced chats and inject into conversation","Observed from Claude Desktop UI showing MCP integration pattern with fetch tool","Would be more robust than passing references in chat text"]}
-{"type":"entity","name":"Project_Priorities","entityType":"roadmap","observations":["P1: Dogfooding basic-memory system instead of JSON memory store","Future: Implement MCP-based reference system"]}
-{"type":"entity","name":"great_observation_loading_saga_20241207","entityType":"debugging_session","observations":["Occurred on December 7, 2024 while debugging basic-memory SQLAlchemy relationship loading","Issue: selectinload() wasn't properly loading relationships in async SQLAlchemy context","Tried multiple solutions: explicit joins, manual loading, various SQLAlchemy loading strategies","Final solution: Using session.refresh() with explicit relationship names","Memorable quote: 'The Great Observation Loading Saga'","Key learning: Sometimes the obvious SQLAlchemy patterns need adaptation for async contexts","Solution preserved in basic-memory repository in EntityRepository.find_by_id()"]}
-{"type":"entity","name":"basic_memory_implementation_20241208","entityType":"technical_milestone","observations":["Fixed async SQLAlchemy relationship loading issues by using explicit refresh with relationship names","Established pattern of relationship handling belonging in MemoryService not EntityService","Fixed ID generation flow through Pydantic schemas to DB layer","Standardized error handling using EntityNotFoundError","All 32 tests passing with 70% coverage","Core services (Entity, Observation, Relation) working properly","Ready for MCP server implementation","Notable debugging session: The Great Observation Loading Saga - resolved lazy loading issues","Established clear separation between MemoryService orchestration and individual service responsibilities"]}
-{"type":"entity","name":"MCP_Dependency_Risk","entityType":"technical_lesson","observations":["Experienced disruption when MCP npm package disappeared - 'leftpad moment'","Need to ensure basic-memory tools are resilient to external dependency issues","Local implementation of MCP server provides better stability than npm packages","Important to maintain control of critical infrastructure components","Validates DIY/local-first philosophy of basic-memory project","Package manager fragility revealed by simple 'npx @modelcontextprotocol/server-memory' failure"]}
-{"type":"entity","name":"basic_memory_project_20241208","entityType":"technical_milestone","observations":["Core MCP server implementation completed with tools: create_entities, search_nodes, open_nodes, add_observations, create_relations, delete_entities, delete_observations","ProjectConfig and dependency injection pattern established","Test framework in place with in-memory DB support","Support for both camelCase (MCP) and snake_case (internal) formats","Filesystem remains source of truth with SQLite as index","Two-way sync pattern identified between Claude MCP tools and direct markdown file editing","Ready for Claude Desktop integration testing phase","Next steps identified: passing tests, markdown format definition, file change tracking, real-world testing","Implementation prioritizes local-first principles with filesystem as source of truth"]}
-{"type":"entity","name":"basic_memory_mcp_architecture","entityType":"technical_design","observations":["MemoryServer class extends MCP Server with custom handler registration","Uses ProjectConfig for clean dependency injection and configuration","Memory service can be injected for testing","Handlers exposed as instance attributes for testing","Tool schemas leverage existing Pydantic models"]}
-{"type":"entity","name":"basic_memory_sync_considerations","entityType":"design_insight","observations":["Need to handle sync between direct markdown file edits and DB index","Watch for file system changes as potential future enhancement","Consider index rebuild patterns on startup","Keep human-friendly markdown format for direct editing"]}
-{"type":"entity","name":"mcp_server_learnings","entityType":"developer_insight","observations":["MCP protocol is new and documentation is still evolving","Test patterns are not well established yet in example implementations","Supporting both camelCase and snake_case helps with protocol/internal compatibility","Server.handle_* naming convention is important for handler registration"]}
-{"type":"entity","name":"20241208-mcp-tool-refactoring","entityType":"conversation","observations":["Decision to return structured data via EmbeddedResource instead of TextContent string parsing","Plan to create Pydantic result models (CreateEntitiesResult, SearchNodesResult etc)","Will use application/vnd.basic-memory+json as MIME type for our structured data","Currently debugging test issues with add_observations tool","Entity ID vs name resolution needed in add_observations","Goal is to make tools more joyful to use by eliminating string parsing","MCP spec supports EmbeddedResource for structured data returns"]}
-{"type":"entity","name":"Basic Memory MCP Server Implementation","entityType":"technical_notes","observations":["Server implements Model Context Protocol using proper structured data responses","Uses EmbeddedResource with custom MIME type 'application/vnd.basic-memory+json'","Clean separation between input validation and handlers via Pydantic models","All tool operations return structured data through create_response helper","Type safety with Literal types for tool names and proper typing for handlers","Handler registry pattern with TOOL_HANDLERS dictionary","Consistent error handling pattern using MCP error codes","Uses Pydantic ConfigDict for proper ORM integration","Tool schemas organized into Input and Response types","Input validation with Annotated types for extra constraints","Response models consistently use from_attributes=True for ORM data","Entity ID generation moved to model validator on EntityBase","Follows principle of making common operations easy and safe"]}
-{"type":"relation","from":"Paul","to":"Basic_Machines","relationType":"created_and_maintains"}
-{"type":"relation","from":"basic-memory","to":"Basic_Machines","relationType":"is_component_of"}
-{"type":"relation","from":"Paul","to":"basic-memory","relationType":"develops"}
-{"type":"relation","from":"fileio_module","to":"basic-memory_implementation_patterns","relationType":"implements"}
-{"type":"relation","from":"entity_service","to":"basic-memory_implementation_patterns","relationType":"implements"}
-{"type":"relation","from":"observation_service","to":"basic-memory_implementation_patterns","relationType":"implements"}
-{"type":"relation","from":"fileio_module","to":"basic-memory","relationType":"is_component_of"}
-{"type":"relation","from":"entity_service","to":"basic-memory","relationType":"is_component_of"}
-{"type":"relation","from":"observation_service","to":"basic-memory","relationType":"is_component_of"}
-{"type":"relation","from":"entity_service","to":"fileio_module","relationType":"uses"}
-{"type":"relation","from":"observation_service","to":"fileio_module","relationType":"uses"}
-{"type":"relation","from":"observation_management","to":"observation_service","relationType":"influences_design_of"}
-{"type":"relation","to":"basic-memory","from":"testing_infrastructure","relationType":"supports"}
-{"type":"relation","to":"testing_infrastructure","from":"test_categories","relationType":"implements"}
-{"type":"relation","to":"basic-memory","from":"completed_work","relationType":"tracks_progress_of"}
-{"type":"relation","to":"basic-memory","from":"future_work","relationType":"guides_development_of"}
-{"type":"relation","to":"basic-memory","from":"design_decisions","relationType":"shapes_architecture_of"}
-{"type":"relation","to":"basic-memory","from":"concurrency_considerations","relationType":"influences_design_of"}
-{"type":"relation","to":"future_work","from":"concurrency_considerations","relationType":"informs"}
-{"type":"relation","to":"observation_management","from":"design_decisions","relationType":"guides"}
-{"type":"relation","to":"testing_infrastructure","from":"completed_work","relationType":"established"}
-{"type":"relation","to":"design_decisions","from":"fileio_module","relationType":"implements"}
-{"type":"relation","from":"observation_update_approaches","to":"observation_management","relationType":"analyzes"}
-{"type":"relation","from":"bulk_update_approach","to":"observation_update_approaches","relationType":"is_option_of"}
-{"type":"relation","from":"tracked_observations_approach","to":"observation_update_approaches","relationType":"is_option_of"}
-{"type":"relation","from":"diff_based_approach","to":"observation_update_approaches","relationType":"is_option_of"}
-{"type":"relation","from":"position_based_approach","to":"observation_update_approaches","relationType":"is_option_of"}
-{"type":"relation","from":"tasks_and_progress","to":"basic-memory","relationType":"tracks_status_of"}
-{"type":"relation","from":"design_decisions","to":"observation_update_approaches","relationType":"influences"}
-{"type":"relation","from":"observation_update_approaches","to":"future_work","relationType":"informs"}
-{"type":"relation","to":"basic-memory_implementation_patterns","from":"error_handling_patterns","relationType":"is_part_of"}
-{"type":"relation","to":"basic-memory","from":"data_models","relationType":"implements"}
-{"type":"relation","to":"basic-memory","from":"markdown_format","relationType":"defines"}
-{"type":"relation","to":"basic-memory","from":"test_driven_development","relationType":"guides_development_of"}
-{"type":"relation","to":"basic-memory","from":"architecture_evolution","relationType":"describes_development_of"}
-{"type":"relation","to":"basic-memory_implementation_patterns","from":"validation_patterns","relationType":"is_part_of"}
-{"type":"relation","to":"design_decisions","from":"architecture_evolution","relationType":"informs"}
-{"type":"relation","to":"fileio_module","from":"markdown_format","relationType":"implements"}
-{"type":"relation","to":"error_handling_patterns","from":"test_driven_development","relationType":"influenced"}
-{"type":"relation","to":"data_models","from":"validation_patterns","relationType":"implements"}
-{"type":"relation","to":"markdown_format","from":"markdown_examples","relationType":"documents"}
-{"type":"relation","to":"markdown_format","from":"markdown_parsing_rules","relationType":"defines"}
-{"type":"relation","to":"data_models","from":"schema_definitions","relationType":"documents"}
-{"type":"relation","to":"test_driven_development","from":"test_evolution","relationType":"describes"}
-{"type":"relation","to":"architecture_evolution","from":"implementation_challenges","relationType":"influenced"}
-{"type":"relation","to":"test_evolution","from":"implementation_challenges","relationType":"shaped"}
-{"type":"relation","to":"future_work","from":"implementation_challenges","relationType":"informs"}
-{"type":"relation","from":"Basic_Factory","to":"Basic_Machines","relationType":"implements"}
-{"type":"relation","from":"Basic_Factory_Components","to":"Basic_Factory","relationType":"is_part_of"}
-{"type":"relation","from":"Component_Translation_Process","to":"Basic_Factory_Components","relationType":"enables"}
-{"type":"relation","from":"Basic_Machines_Philosophy","to":"Basic_Machines","relationType":"guides"}
-{"type":"relation","from":"Paul","to":"Basic_Factory","relationType":"develops"}
-{"type":"relation","from":"Paul","to":"Basic_Machines_Philosophy","relationType":"created"}
-{"type":"relation","from":"Basic_Machines_Manifesto","to":"Basic_Machines_Philosophy","relationType":"articulates"}
-{"type":"relation","from":"AI_Human_Collaboration_Model","to":"Basic_Factory","relationType":"guides_development_of"}
-{"type":"relation","from":"Basic_Machines_Manifesto","to":"Paul","relationType":"written_by"}
-{"type":"relation","from":"Basic_Machines_Manifesto","to":"Component_Translation_Process","relationType":"documents"}
-{"type":"relation","from":"AI_Human_Collaboration_Model","to":"Basic_Machines","relationType":"shapes_development_of"}
-{"type":"relation","from":"Basic_Machines_Roadmap","to":"Basic_Machines","relationType":"guides_development_of"}
-{"type":"relation","from":"Basic_Machines_Website","to":"Basic_Machines_Roadmap","relationType":"implements_phase_of"}
-{"type":"relation","from":"Basic_Factory_Components","to":"Basic_Machines_Website","relationType":"enables"}
-{"type":"relation","from":"Basic_Machines_Philosophy","to":"Basic_Machines_Website","relationType":"informs"}
-{"type":"relation","from":"Paul","to":"DIY_Ethics","relationType":"embodies"}
-{"type":"relation","from":"Basic_Machines_Philosophy","to":"DIY_Ethics","relationType":"incorporates"}
-{"type":"relation","from":"Basic_Machines","to":"DIY_Ethics","relationType":"exemplifies"}
-{"type":"relation","from":"Component_Translation_Process","to":"Basic_Machines_Philosophy","relationType":"implements"}
-{"type":"relation","from":"Basic_Factory_Components","to":"DIY_Ethics","relationType":"demonstrates"}
-{"type":"relation","from":"AI_Human_Collaboration_Model","to":"Basic_Machines_Philosophy","relationType":"aligns_with"}
-{"type":"relation","from":"AI_Human_Collaboration_Model","to":"Component_Translation_Process","relationType":"guides"}
-{"type":"relation","from":"Basic_Machines_Manifesto","to":"Basic_Machines","relationType":"defines_vision_for"}
-{"type":"relation","from":"Basic_Machines_Website","to":"Basic_Machines_Manifesto","relationType":"implements_vision_of"}
-{"type":"relation","from":"Basic_Factory","to":"AI_Human_Collaboration_Model","relationType":"demonstrates"}
-{"type":"relation","from":"Paul","to":"AI_Human_Collaboration_Model","relationType":"developed_with_Claude"}
-{"type":"relation","from":"Basic_Factory_Components","to":"Component_Translation_Process","relationType":"created_through"}
-{"type":"relation","from":"Basic_Machines_Philosophy","to":"Basic_Factory","relationType":"guides"}
-{"type":"relation","from":"Basic_Factory","to":"MCP_Tools","relationType":"integrates"}
-{"type":"relation","from":"Basic_Machines_Website","to":"Basic_Factory_Components","relationType":"will_use"}
-{"type":"relation","from":"Basic_Machines_Roadmap","to":"Basic_Machines_Philosophy","relationType":"aligns_with"}
-{"type":"relation","from":"Component_Translation_Process","to":"MCP_Tools","relationType":"leverages"}
-{"type":"relation","from":"Basic_Factory","to":"basic-memory","relationType":"will_document_process_in"}
-{"type":"relation","from":"AI_Human_Collaboration_Model","to":"basic-memory","relationType":"will_be_implemented_in"}
-{"type":"relation","from":"Basic_Machines_Philosophy","to":"Basic_Machines_Roadmap","relationType":"informs_priorities_of"}
-{"type":"relation","from":"basic-memory","to":"Basic_Machines_Philosophy","relationType":"embodies"}
-{"type":"relation","from":"Paul","to":"Basic_Machines_Manifesto","relationType":"authored_with_Claude"}
-{"type":"relation","from":"Basic_Factory","to":"Component_Translation_Process","relationType":"validated"}
-{"type":"relation","from":"AI_Human_Collaboration_Model","to":"MCP_Tools","relationType":"utilizes"}
-{"type":"relation","from":"Basic_Factory_Components","to":"Basic_Machines_Roadmap","relationType":"supports"}
-{"type":"relation","from":"Basic_Machines_Website","to":"Basic_Factory","relationType":"will_demonstrate"}
-{"type":"relation","from":"Basic_Factory","to":"basic-memory-webui","relationType":"enables_development_of"}
-{"type":"relation","from":"basic-memory","to":"AI_Human_Development_Methodology","relationType":"implements"}
-{"type":"relation","from":"Basic_Machines_Philosophy","to":"AI_Human_Development_Methodology","relationType":"guides"}
-{"type":"relation","from":"Basic_Factory_Components","to":"basic-memory-webui","relationType":"provides_ui_for"}
-{"type":"relation","from":"Component_Translation_Process","to":"AI_Human_Development_Methodology","relationType":"exemplifies"}
-{"type":"relation","from":"Basic_Factory","to":"Basic Components","relationType":"enabled_creation_of"}
-{"type":"relation","from":"Basic_Factory","to":"Tool Integration Discovery","relationType":"led_to"}
-{"type":"relation","from":"MCP_Integration_Progress","to":"AI_Human_Development_Methodology","relationType":"validates"}
-{"type":"relation","from":"Basic_Factory","to":"MCP_Integration_Progress","relationType":"demonstrates"}
-{"type":"relation","from":"Basic_Factory","to":"AI_Human_Development_Methodology","relationType":"proves_effectiveness_of"}
-{"type":"relation","from":"Basic_Machines_Philosophy","to":"Basic Components","relationType":"inspires_architecture_of"}
-{"type":"relation","from":"DIY_Ethics","to":"basic-memory","relationType":"shapes_design_of"}
-{"type":"relation","from":"Basic_Machines_Philosophy","to":"Tool Integration Discovery","relationType":"guides_analysis_of"}
-{"type":"relation","from":"Basic_Machines_Manifesto","to":"AI_Human_Development_Methodology","relationType":"documents_approach_of"}
-{"type":"relation","from":"Basic_Machines_Manifesto","to":"Basic_Factory_Components","relationType":"explains_principles_of"}
-{"type":"relation","from":"Basic_Memory_Project_Structure","to":"basic-memory","relationType":"organizes"}
-{"type":"relation","from":"Basic_Memory_Database_Schema","to":"basic-memory","relationType":"defines_storage_for"}
-{"type":"relation","from":"Basic_Memory_Markdown_Example","to":"Basic_Memory_File_Format","relationType":"demonstrates"}
-{"type":"relation","from":"Basic_Memory_Project_Isolation_Decision","to":"Basic_Memory_Future_Enhancement_Weighted_Relations","relationType":"similar_to"}
-{"type":"relation","to":"DIY_Ethics","from":"Basic_Memory_Project_Isolation_Decision","relationType":"follows"}
-{"type":"relation","from":"Basic_Memory_Implementation_Plan","to":"basic-memory","relationType":"guides"}
-{"type":"relation","from":"Basic_Memory_Implementation_Plan","to":"DIY_Ethics","relationType":"follows"}
-{"type":"relation","from":"Basic_Memory_Implementation_Plan","to":"Basic_Memory_Database_Schema","relationType":"implements"}
-{"type":"relation","from":"Basic_Memory_Implementation_Status","to":"Basic_Memory_Implementation_Plan","relationType":"updates"}
-{"type":"relation","from":"Basic_Memory_Observation_Management_Design","to":"Basic_Memory_Technical_Design","relationType":"extends"}
-{"type":"relation","from":"Basic_Memory_Architectural_Decisions","to":"DIY_Ethics","relationType":"guided_by"}
-{"type":"relation","from":"Basic_Memory_Architectural_Decisions","to":"basic-memory","relationType":"structures"}
-{"type":"relation","from":"Basic_Memory_Implementation_Status","to":"basic-memory","relationType":"describes_state_of"}
-{"type":"relation","to":"Basic_Memory_Implementation_Status","from":"Basic_Memory_Implementation_Analysis","relationType":"analyzes"}
-{"type":"relation","to":"basic-memory","from":"Basic_Memory_Current_Challenges","relationType":"identifies_issues_in"}
-{"type":"relation","to":"DIY_Ethics","from":"Basic_Memory_Implementation_Analysis","relationType":"confirms_alignment_with"}
-{"type":"relation","to":"Basic_Memory_Observation_Management_Design","from":"Basic_Memory_Observation_Hash_Tracking","relationType":"solves"}
-{"type":"relation","to":"DIY_Ethics","from":"Basic_Memory_Observation_Hash_Tracking","relationType":"aligns_with"}
-{"type":"relation","to":"Basic_Memory_File_Format","from":"Basic_Memory_Observation_Hash_Tracking","relationType":"preserves"}
-{"type":"relation","to":"Basic_Memory_Technical_Design","from":"Basic_Memory_Observation_Hash_Tracking","relationType":"enhances"}
-{"type":"relation","from":"Basic_Memory_Repository_Implementation","to":"basic-memory","relationType":"implements_part_of"}
-{"type":"relation","from":"Basic_Memory_Repository_Implementation","to":"Basic_Memory_Database_Schema","relationType":"follows"}
-{"type":"relation","from":"Basic_Memory_Repository_Implementation","to":"DIY_Ethics","relationType":"aligns_with"}
-{"type":"relation","from":"Basic_Memory_Repository_Implementation","to":"testing_infrastructure","relationType":"demonstrates"}
-{"type":"relation","from":"Basic_Memory_Repository_Implementation","to":"Basic Foundation","relationType":"inspired_by"}
-{"type":"relation","from":"Basic_Memory_Dependencies","to":"basic-memory","relationType":"supports"}
-{"type":"relation","from":"Basic_Memory_Dependencies","to":"Basic_Memory_Repository_Implementation","relationType":"enables"}
-{"type":"relation","from":"Basic_Memory_Dependencies","to":"testing_infrastructure","relationType":"enables"}
-{"type":"relation","from":"Basic_Memory_Current_Architecture","to":"basic-memory","relationType":"describes_state_of"}
-{"type":"relation","from":"Basic_Memory_Evolution","to":"Basic_Memory_Current_Architecture","relationType":"explains_development_of"}
-{"type":"relation","from":"Basic_Memory_Service_Layer","to":"Basic_Memory_Current_Architecture","relationType":"implements"}
-{"type":"relation","from":"Basic_Memory_Schema_Design","to":"Basic_Memory_Current_Architecture","relationType":"implements"}
-{"type":"relation","from":"Basic_Memory_Evolution","to":"Basic_Memory_Implementation_Plan","relationType":"reflects_on"}
-{"type":"relation","from":"Basic_Memory_Evolution","to":"DIY_Ethics","relationType":"demonstrates_alignment_with"}
-{"type":"relation","from":"Basic_Memory_Current_Architecture","to":"DIY_Ethics","relationType":"embodies"}
-{"type":"relation","from":"Basic_Memory_Service_Layer","to":"fileio_module","relationType":"uses"}
-{"type":"relation","from":"Basic_Memory_Schema_Design","to":"markdown_format","relationType":"implements"}
-{"type":"relation","to":"basic-memory","from":"Basic_Memory_Next_Tasks","relationType":"guides_development_of"}
-{"type":"relation","to":"DIY_Ethics","from":"Basic_Memory_Next_Tasks","relationType":"aligns_with"}
-{"type":"relation","to":"Basic_Memory_Current_Architecture","from":"Basic_Memory_Next_Tasks","relationType":"extends"}
-{"type":"relation","from":"Basic_Memory_Meta_Experience","to":"basic-memory","relationType":"validates_design_of"}
-{"type":"relation","from":"Basic_Memory_Meta_Experience","to":"DIY_Ethics","relationType":"demonstrates_principles_of"}
-{"type":"relation","from":"Basic_Memory_Meta_Experience","to":"design_decisions","relationType":"reinforces"}
-{"type":"relation","from":"Basic_Memory_Meta_Experience","to":"Basic_Memory_Current_Architecture","relationType":"validates"}
-{"type":"relation","from":"Model_Context_Protocol","to":"basic-memory","relationType":"enables"}
-{"type":"relation","from":"basic-memory_core_principles","to":"basic-memory","relationType":"guides"}
-{"type":"relation","from":"basic-memory_core_principles","to":"DIY_Ethics","relationType":"aligns_with"}
-{"type":"relation","from":"basic-memory_business_model","to":"basic-memory","relationType":"defines_sustainability_for"}
-{"type":"relation","from":"basic-memory_business_model","to":"DIY_Ethics","relationType":"maintains_alignment_with"}
-{"type":"relation","from":"basic-memory_cli","to":"basic-memory","relationType":"provides_interface_for"}
-{"type":"relation","from":"basic-memory_cli","to":"Model_Context_Protocol","relationType":"integrates_with"}
-{"type":"relation","from":"basic-memory_export_format","to":"basic-memory","relationType":"standardizes_output_of"}
-{"type":"relation","from":"basic-memory_export_format","to":"markdown_format","relationType":"extends"}
-{"type":"relation","from":"basic-memory_core_principles","to":"Basic_Machines_Philosophy","relationType":"implements"}
-{"type":"relation","from":"Model_Context_Protocol","to":"AI_Human_Collaboration_Model","relationType":"enables"}
-{"type":"relation","from":"relation_service","to":"basic-memory","relationType":"will_be_component_of"}
-{"type":"relation","from":"relation_service","to":"service_layer_patterns","relationType":"follows"}
-{"type":"relation","from":"relation_service","to":"fileio_patterns","relationType":"uses"}
-{"type":"relation","from":"relation_service","to":"database_models","relationType":"uses"}
-{"type":"relation","from":"relation_service","to":"repository_patterns","relationType":"implements"}
-{"type":"relation","from":"relation_service_design","to":"relation_service","relationType":"guides_implementation_of"}
-{"type":"relation","from":"relation_service_implementation_plan","to":"relation_service","relationType":"defines_implementation_of"}
-{"type":"relation","from":"relation_service_challenges","to":"relation_service_design","relationType":"informs"}
-{"type":"relation","from":"relation_file_format","to":"markdown_format","relationType":"extends"}
-{"type":"relation","from":"relation_service_error_handling","to":"service_layer_patterns","relationType":"implements"}
-{"type":"relation","from":"relation_service_testing","to":"testing_infrastructure","relationType":"extends"}
-{"type":"relation","from":"fileio_patterns","to":"service_layer_patterns","relationType":"enables"}
-{"type":"relation","from":"database_models","to":"repository_patterns","relationType":"enables"}
-{"type":"relation","from":"relation_service","to":"entity_service","relationType":"coordinates_with"}
-{"type":"relation","from":"relation_file_format","to":"relation_service","relationType":"defines_storage_for"}
-{"type":"relation","from":"relation_service_error_handling","to":"relation_service","relationType":"ensures_reliability_of"}
-{"type":"relation","from":"relation_service_testing","to":"relation_service","relationType":"verifies"}
-{"type":"relation","from":"service_layer_patterns","to":"basic-memory_implementation_patterns","relationType":"implements"}
-{"type":"relation","from":"repository_patterns","to":"basic-memory_implementation_patterns","relationType":"implements"}
-{"type":"relation","from":"fileio_patterns","to":"basic-memory_implementation_patterns","relationType":"implements"}
-{"type":"relation","from":"database_models","to":"basic-memory_implementation_patterns","relationType":"implements"}
-{"type":"relation","from":"relation_service_challenges","to":"implementation_challenges","relationType":"extends"}
-{"type":"relation","from":"relation_service_implementation_plan","to":"future_work","relationType":"details"}
-{"type":"relation","from":"relation_service_design","to":"design_decisions","relationType":"aligns_with"}
-{"type":"relation","from":"relation_file_format","to":"design_decisions","relationType":"follows"}
-{"type":"relation","from":"pytest_patterns","to":"testing_infrastructure","relationType":"extends"}
-{"type":"relation","from":"relation_implementation_learnings","to":"basic-memory_implementation_patterns","relationType":"informs"}
-{"type":"relation","from":"test_driven_insights","to":"test_driven_development","relationType":"enriches"}
-{"type":"relation","from":"meta_development_insights","to":"AI_Human_Collaboration_Model","relationType":"improves"}
-{"type":"relation","from":"relation_implementation_learnings","to":"relation_service","relationType":"guides_implementation_of"}
-{"type":"relation","from":"pytest_patterns","to":"test_evolution","relationType":"demonstrates"}
-{"type":"relation","from":"test_driven_insights","to":"design_decisions","relationType":"influences"}
-{"type":"relation","from":"meta_development_insights","to":"architecture_evolution","relationType":"informs"}
-{"type":"relation","from":"relation_service","to":"relation_implementation_learnings","relationType":"validates"}
-{"type":"relation","from":"test_driven_insights","to":"implementation_challenges","relationType":"helps_solve"}
-{"type":"relation","from":"AI_Assistant_Learnings","to":"meta_development_insights","relationType":"enriches"}
-{"type":"relation","from":"Effective_Response_Patterns","to":"AI_Assistant_Learnings","relationType":"implements"}
-{"type":"relation","from":"AI_Context_Management","to":"AI_Human_Collaboration_Model","relationType":"improves"}
-{"type":"relation","from":"AI_Tool_Usage_Patterns","to":"AI_Context_Management","relationType":"enables"}
-{"type":"relation","from":"AI_Assistant_Learnings","to":"Basic_Memory_Meta_Experience","relationType":"validates"}
-{"type":"relation","from":"AI_Tool_Usage_Patterns","to":"Model_Context_Protocol","relationType":"demonstrates_effective_use_of"}
-{"type":"relation","from":"AI_Context_Management","to":"basic-memory","relationType":"validates_design_of"}
-{"type":"relation","from":"Effective_Response_Patterns","to":"AI_Human_Development_Methodology","relationType":"refines"}
-{"type":"relation","to":"relation_service","from":"relation_service_patterns","relationType":"guides"}
-{"type":"relation","to":"test_driven_development","from":"test_driven_insights_relations","relationType":"enriches"}
-{"type":"relation","to":"implementation_challenges","from":"relation_service_learnings","relationType":"solves"}
-{"type":"relation","to":"basic-memory_implementation_patterns","from":"relation_service_patterns","relationType":"implements"}
-{"type":"relation","to":"markdown_format","from":"relation_service_patterns","relationType":"extends"}
-{"type":"relation","to":"service_layer_patterns","from":"relation_service_patterns","relationType":"refines"}
-{"type":"relation","from":"packaging_learnings","to":"implementation_challenges","relationType":"informs"}
-{"type":"relation","from":"packaging_learnings","to":"test_driven_development","relationType":"impacts"}
-{"type":"relation","to":"basic-memory","from":"Recent_Implementation_Progress","relationType":"updates_status_of"}
-{"type":"relation","to":"future_work","from":"Next_Steps","relationType":"extends"}
-{"type":"relation","to":"design_decisions","from":"Development_Practices","relationType":"informs"}
-{"type":"relation","to":"packaging_learnings","from":"Development_Practices","relationType":"incorporates"}
-{"type":"relation","to":"test_driven_development","from":"Development_Practices","relationType":"refines"}
-{"type":"relation","to":"basic-memory_implementation_patterns","from":"Development_Practices","relationType":"enhances"}
-{"type":"relation","from":"Basic_Memory_MCP","to":"MCP_Server_Implementation","relationType":"follows"}
-{"type":"relation","from":"Basic_Memory_MCP","to":"MCP_Tools","relationType":"uses"}
-{"type":"relation","from":"Basic_Memory","to":"MCP_Server_Implementation","relationType":"implements"}
-{"type":"relation","from":"Basic_Memory_Testing","to":"Memory_Service_Tests","relationType":"includes"}
-{"type":"relation","from":"Basic_Memory_Testing","to":"MCP_Server_Tests","relationType":"includes"}
-{"type":"relation","from":"Memory_Service_Tests","to":"Basic_Memory_MCP","relationType":"validates"}
-{"type":"relation","from":"MCP_Server_Tests","to":"Basic_Memory_MCP","relationType":"validates"}
-{"type":"relation","from":"Service_Interface_Audit","to":"Memory_Service_Refactoring","relationType":"informs"}
-{"type":"relation","from":"Memory_Service_Refactoring","to":"Basic_Memory_MCP","relationType":"affects"}
-{"type":"relation","from":"Memory_Service_Patterns","to":"Basic_Memory_MCP","relationType":"improves"}
-{"type":"relation","from":"Pydantic_Create_Pattern","to":"Memory_Service_Patterns","relationType":"enables"}
-{"type":"relation","from":"Pydantic_Create_Pattern","to":"Basic_Memory_MCP","relationType":"improves"}
-{"type":"relation","from":"MCP_Marketplace","to":"Basic_Memory_Business","relationType":"enables"}
-{"type":"relation","from":"Basic_Memory","to":"MCP_Marketplace","relationType":"could_integrate_with"}
-{"type":"relation","from":"Persistence_Of_Vision","to":"Basic_Memory","relationType":"helps_achieve"}
-{"type":"relation","from":"Drew","to":"Persistence_Of_Vision","relationType":"conceptualized"}
-{"type":"relation","to":"Usage_Recipes","from":"Conversation_Continuity_Pattern","relationType":"is_example_of"}
-{"type":"relation","to":"Basic_Memory","from":"Usage_Recipes","relationType":"enhances"}
-{"type":"relation","to":"Basic_Memory","from":"Chat_References","relationType":"enhances"}
-{"type":"relation","to":"Conversation_Continuity_Pattern","from":"Chat_References","relationType":"implements"}
-{"type":"relation","from":"20240307-chat-reference-protocol-test","to":"20240307-chat-reference-protocol","relationType":"continues_from"}
-{"type":"relation","from_id":"Run_Tests_Tool_Request","to_id":"Basic_Machines","relationType":"enhances","context":"development workflow improvement"}
-{"type":"relation","from_id":"SQLAlchemy_Async_Loading_Pattern","to_id":"basic-memory","relation_type":"improves","context":"database performance and async compatibility"}
-{"type":"relation","from_id":"SQLAlchemy_Async_Loading_Pattern","to_id":"Entity","relation_type":"applies_to","context":"relationship loading strategy"}
-{"type":"relation","from":"MCP_Reference_Integration","to":"Project_Priorities","relationType":"prioritized_after"}
-{"type":"relation","from":"great_observation_loading_saga_20241207","to":"Basic_Memory","relationType":"occurred_in"}
-{"type":"relation","from":"great_observation_loading_saga_20241207","to":"SQLAlchemy","relationType":"relates_to"}
-{"type":"relation","from":"basic_memory_implementation_20241208","to":"Basic_Memory","relationType":"improves"}
-{"type":"relation","from":"great_observation_loading_saga_20241207","to":"basic_memory_implementation_20241208","relationType":"leads_to"}
-{"type":"relation","from":"MCP_Dependency_Risk","to":"DIY_Ethics","relationType":"validates"}
-{"type":"relation","from":"MCP_Dependency_Risk","to":"basic-memory_core_principles","relationType":"reinforces"}
-{"type":"relation","from":"MCP_Dependency_Risk","to":"Basic_Memory_Implementation_Plan","relationType":"influences"}
-{"type":"relation","from":"basic_memory_mcp_architecture","to":"basic_memory_project_20241208","relationType":"implements"}
-{"type":"relation","from":"basic_memory_sync_considerations","to":"basic_memory_project_20241208","relationType":"influences"}
-{"type":"relation","from":"mcp_server_learnings","to":"basic_memory_mcp_architecture","relationType":"informs"}
-{"type":"relation","from":"20241208-mcp-tool-refactoring","to":"Basic_Memory_MCP","relationType":"improves"}
-{"type":"relation","from":"20241208-mcp-tool-refactoring","to":"Basic_Memory_Implementation_Plan","relationType":"implements"}

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 = ">=3.12.1"
+requires-python = ">=3.12"
 license = { text = "AGPL-3.0-or-later" }
 authors = [
     { name = "Basic Machines", email = "hello@basic-machines.co" }
@@ -14,9 +14,8 @@ dependencies = [
     "typer>=0.9.0",
     "aiosqlite>=0.20.0",
     "greenlet>=3.1.1",
-    "pydantic[email,timezone]>=2.10.3",
-    "icecream>=2.1.3",
-    "mcp>=1.2.0",
+    "pydantic[email,timezone]>=2.12.0",
+    "mcp>=1.23.1",
     "pydantic-settings>=2.6.1",
     "loguru>=0.7.3",
     "pyright>=1.1.390",
@@ -30,12 +29,29 @@ dependencies = [
     "alembic>=1.14.1",
     "pillow>=11.1.0",
     "pybars3>=0.9.7",
-    "fastmcp>=2.3.4",
+    "fastmcp==2.12.3", # Pinned - 2.14.x breaks MCP tools visibility (issue #463)
     "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",
 ]
 
+[project.optional-dependencies]
+semantic = [
+    "fastembed>=0.7.4",
+    "sqlite-vec>=0.1.6",
+    "openai>=1.100.2",
+]
 
 [project.urls]
 Homepage = "https://github.com/basicmachines-co/basic-memory"
@@ -52,17 +68,24 @@ build-backend = "hatchling.build"
 
 [tool.pytest.ini_options]
 pythonpath = ["src", "tests"]
-addopts = "--cov=basic_memory --cov-report term-missing -ra -q"
-testpaths = ["tests"]
+addopts = "--cov=basic_memory --cov-report term-missing"
+testpaths = ["tests", "test-int"]
 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",
+]
 
 [tool.ruff]
 line-length = 100
 target-version = "py312"
 
-[tool.uv]
-dev-dependencies = [
+[dependency-groups]
+dev = [
     "gevent>=24.11.1",
     "icecream>=2.1.3",
     "pytest>=8.3.4",
@@ -71,6 +94,11 @@ dev-dependencies = [
     "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",
 ]
 
 [tool.hatch.version]
@@ -95,6 +123,8 @@ pythonVersion = "3.12"
 
 [tool.coverage.run]
 concurrency = ["thread", "gevent"]
+parallel = true
+source = ["basic_memory"]
 
 [tool.coverage.report]
 exclude_lines = [
@@ -116,11 +146,9 @@ 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/main.py",               # CLI entry point
-    "*/mcp/tools/project_management.py",  # Covered by integration tests
-    "*/mcp/tools/sync_status.py",  # Covered by integration tests
+    "*/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
     "*/services/migration_service.py", # Complex migration scenarios
 ]
-
-[tool.logfire]
-ignore_no_config = true

scripts/install.sh

@@ -1,36 +0,0 @@
-#!/bin/bash
-set -e
-
-echo "Welcome to Basic Memory installer"
-
-# 1. Install uv if not present
-if ! command -v uv &> /dev/null; then
-    echo "Installing uv package manager..."
-    curl -LsSf https://github.com/astral-sh/uv/releases/download/0.1.23/uv-installer.sh | sh
-fi
-
-# 2. Configure Claude Desktop
-echo "Configuring Claude Desktop..."
-CONFIG_FILE="$HOME/Library/Application Support/Claude/claude_desktop_config.json"
-
-# Create config directory if it doesn't exist
-mkdir -p "$(dirname "$CONFIG_FILE")"
-
-# If config file doesn't exist, create it with initial structure
-if [ ! -f "$CONFIG_FILE" ]; then
-    echo '{"mcpServers": {}}' > "$CONFIG_FILE"
-fi
-
-# Add/update the basic-memory config using jq
-jq '.mcpServers."basic-memory" = {
-    "command": "uvx",
-    "args": ["basic-memory"]
-}' "$CONFIG_FILE" > "$CONFIG_FILE.tmp" && mv "$CONFIG_FILE.tmp" "$CONFIG_FILE"
-
-echo "Installation complete! Basic Memory is now available in Claude Desktop."
-echo "Please restart Claude Desktop for changes to take effect."
-
-echo -e "\nQuick Start:"
-echo "1. You can run sync directly using: uvx basic-memory sync"
-echo "2. Optionally, install globally with: uv pip install basic-memory"
-echo -e "\nBuilt with ♥️ by Basic Machines."

server.json

@@ -0,0 +1,25 @@
+{
+  "$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": "0.18.3",
+  "packages": [
+    {
+      "registryType": "pypi",
+      "identifier": "basic-memory",
+      "version": "0.18.3",
+      "runtimeHint": "uvx",
+      "runtimeArguments": [
+        {"type": "positional", "value": "basic-memory"},
+        {"type": "positional", "value": "mcp"}
+      ],
+      "transport": {
+        "type": "stdio"
+      }
+    }
+  ]
+}

src/basic_memory/__init__.py

@@ -1,4 +1,7 @@
 """basic-memory - Local-first knowledge management combining Zettelkasten with knowledge graphs"""
 
+# Package version - updated by release automation
+__version__ = "0.18.3"
+
 # API version for FastAPI - independent of package version
-__version__ = "v0"
+__api_version__ = "v0"

src/basic_memory/alembic/env.py

@@ -1,29 +1,60 @@
 """Alembic environment configuration."""
 
+import asyncio
 import os
 from logging.config import fileConfig
 
-from sqlalchemy import engine_from_config
-from sqlalchemy import pool
+# 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 alembic import context
 
-from basic_memory.models import Base
+from basic_memory.config import ConfigManager
 
-# set config.env to "test" for pytest to prevent logging to file in utils.setup_logging()
-os.environ["BASIC_MEMORY_ENV"] = "test"
+# 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"
 
-from basic_memory.config import app_config
+# Import after setting environment variable  # 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
 
-# Set the SQLAlchemy URL from our app config
-sqlalchemy_url = f"sqlite:///{app_config.database_path}"
-config.set_main_option("sqlalchemy.url", sqlalchemy_url)
+# 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
 
-# print(f"Using 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)
 
 # Interpret the config file for Python logging.
 if config.config_file_name is not None:
@@ -67,28 +98,89 @@ 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.
 
-    In this scenario we need to create an Engine
-    and associate a connection with the context.
+    Supports both sync engines (SQLite) and async engines (PostgreSQL with asyncpg).
     """
-    connectable = engine_from_config(
-        config.get_section(config.config_ini_section, {}),
-        prefix="sqlalchemy.",
-        poolclass=pool.NullPool,
-    )
+    # Check if a connection/engine was provided (e.g., from run_migrations)
+    connectable = context.config.attributes.get("connection", None)
 
-    with connectable.connect() as connection:
-        context.configure(
-            connection=connection,
-            target_metadata=target_metadata,
-            include_object=include_object,
-            render_as_batch=True,
-        )
+    if connectable is None:
+        # No connection provided, create engine from config
+        url = context.config.get_main_option("sqlalchemy.url")
 
-        with context.begin_transaction():
-            context.run_migrations()
+        # 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)
 
 
 if context.is_offline_mode():

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

@@ -0,0 +1,131 @@
+"""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

@@ -21,6 +21,12 @@ 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),
@@ -55,7 +61,9 @@ 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"),
+            sqlite_where=sa.text("content_type = 'text/markdown' AND permalink IS NOT NULL")
+            if is_sqlite
+            else None,
         )
         batch_op.drop_index("ix_entity_file_path")
         batch_op.create_index(batch_op.f("ix_entity_file_path"), ["file_path"], unique=False)
@@ -67,12 +75,16 @@ 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"),
+            sqlite_where=sa.text("content_type = 'text/markdown' AND permalink IS NOT NULL")
+            if is_sqlite
+            else None,
         )
         batch_op.create_foreign_key("fk_entity_project_id", "project", ["project_id"], ["id"])
 
     # drop the search index table. it will be recreated
-    op.drop_table("search_index")
+    # Only drop for SQLite - Postgres uses regular table managed by ORM
+    if is_sqlite:
+        op.drop_table("search_index")
 
     # ### end Alembic commands ###
 

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

@@ -25,43 +25,51 @@ 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.
 
-    Since SQLite doesn't support dropping specific constraints easily, we'll
-    recreate the table without the problematic constraint.
+    SQLite: Recreate the table without the constraint (no ALTER TABLE support)
+    Postgres: Use ALTER TABLE to drop the constraint directly
     """
-    # 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"),
-    )
+    connection = op.get_bind()
+    is_sqlite = connection.dialect.name == "sqlite"
 
-    # Copy data from old table to new table
-    op.execute("INSERT INTO project_new SELECT * FROM project")
+    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"),
+        )
 
-    # Drop the old table
-    op.drop_table("project")
+        # Copy data from old table to new table
+        op.execute("INSERT INTO project_new SELECT * FROM project")
 
-    # Rename the new table
-    op.rename_table("project_new", "project")
+        # Drop the old table
+        op.drop_table("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)
+        # 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")
 
 
 def downgrade() -> None:

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

@@ -0,0 +1,24 @@
+"""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

@@ -0,0 +1,49 @@
+"""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

@@ -0,0 +1,49 @@
+"""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")