Your development memory across AI sessions.
TaskTracker is a lightweight CLI tool that maintains context about your development work. It's designed for developers who use AI assistants and need to preserve project context across different sessions, tools, and time.
Every developer using AI assistants knows this pain:
- Start new chat β Explain entire project context again
- Switch AI tools β Lose all conversation history
- Take a break β Forget where you left off
- Debug an issue β Can't remember what you already tried
TaskTracker solves this by being your persistent development memory.
# Install globally
npm install -g @dvc2/tasktracker-cli
# Initialize in your project
cd your-project
tt init
# Set your project vision
tt prd "Building a REST API for user management with JWT auth"
# Start documenting your work
tt j "Implemented user registration endpoint"
tt j --type decision "Using bcrypt for password hashing"
tt j --type blocker "JWT refresh token logic is complex"
# Generate context for your AI
tt c # Quick context
tt cf # Full context with historySeamlessly capture your development history:
# Install git hooks for auto-journaling
tt git-install-hook
# Now every commit automatically creates a journal entry!
git commit -m "feat: Add user authentication"
# β Automatically journaled with tags: git, feat, main
# Import recent commits
tt git-sync 10
# Toggle auto-prompting
tt git-auto off # Disable context promptingTrack your progress, decisions, and blockers:
# Quick aliases for common entries
tt done "Completed user authentication flow"
tt decided "Switching from MongoDB to PostgreSQL"
tt blocked "CORS issues with frontend"
tt til "Redis connection pooling improves performance"
# Or use the full command
tt j "Added rate limiting to login endpoint" --tags api,authParse and maintain your project requirements:
tt prd "Build a task management API with real-time updates"
tt prd requirements.md # Or from a file
tt prd-show # View parsed requirementsGenerate rich context for any AI assistant:
tt c # Quick context (last day)
tt cf # Full context (last 7 days)
tt cf 14 # Custom timeframe
tt cf --output ctx.md # Save to fileFind and export your development history:
tt journal-search "authentication" # Search entries
tt journal-show --type decision # Filter by type
tt journal-export markdown # Export journal# Get back up to speed
tt c
# Copy output to your AI assistant# Track progress
tt done "Added user profile endpoints"
# Document decisions
tt decided "Using Redis for session storage - built-in expiration"
# Note blockers
tt blocked "WebSocket connection drops after 30 seconds"# Document the issue
tt blocked "Users can't login - 401 errors"
# Add context
tt j "Checked: JWT secret is correct, token format is valid"
tt j "Suspecting: Token expiration or timezone issue"
# Get focused context for AI
tt cf 1 # Just today's contexttt journal "text"(alias:tt j) - Add entrytt journal-show(alias:tt js) - Show entriestt journal-search "query"- Search entriestt journal-export [format]- Export journal
tt done "text"- Quick progress entrytt decided "text"- Quick decision entrytt blocked "text"- Quick blocker entrytt til "text"- Quick learning entry
tt git-install-hook- Install auto-journaling hooktt git-sync [count]- Import recent commitstt git-auto [on|off]- Toggle auto-promptingtt git-status- Show integration status
tt context-quick(alias:tt c) - Quick contexttt context-full [days](alias:tt cf) - Full context
tt prd "description"- Set project requirementstt prd-show- View current PRDtt prd-context- Generate PRD context
tt init- Initialize TaskTrackertt stats- Show project statisticstt help [command]- Get help
tt j "text" --type [progress|decision|blocker|idea|context]
tt j "text" --tags tag1,tag2,tag3
tt j "text" --files file1.js,file2.pytt journal-show --type decision
tt journal-show --tag architecture
tt journal-show --date 2024-01-15tt cf --output context.md
tt journal-export json --output backup.json- Git Integration - Automatically captures your development history
- AI-First Design - Built specifically for AI-assisted development
- Zero Friction - Simple commands that fit your workflow
- Local & Private - Your data stays in your project
See AI Integration Guide for detailed patterns and workflows with:
- Cursor
- ChatGPT / Claude
- GitHub Copilot
- Custom integrations
- Be Specific: "Fixed null check in auth middleware" > "fixed bug"
- Document Why: Include reasoning in decisions
- Tag Consistently: Use tags to group related work
- Regular Context: Regenerate context at session start
Version 3.0 - Recently refactored from a task manager to a focused developer context journal.
- β All journal functionality (add, show, search, export)
- β PRD parsing and management
- β AI context generation (quick & full)
- β Git integration with auto-journaling
- β Productivity aliases for quick entries
- β Clean, maintainable codebase
- β Comprehensive test coverage
- No cloud sync (by design - local only)
- No team features (focused on individual developers)
- Basic search (no fuzzy matching yet)
- Limited to CLI (no GUI planned)
TaskTracker stores all data locally in the .tasktracker/ directory. This directory contains your development journal and should never be committed to version control.
- Your journal entries are stored locally and privately
- The
.gitignoreautomatically excludes.tasktracker/directories - Never commit real journal data to public repositories
- Use the sanitized examples in
examples/for demos or documentation
Check out the examples directory for:
- Sample journal entries
- Example configurations
- Usage patterns
- Best practices
We welcome contributions! See CONTRIBUTING.md for:
- Development setup
- Code style guide
- Pull request process
- Testing requirements
TaskTracker is MIT licensed.
Built by developers, for developers who use AI assistants.
Special thanks to all contributors who help make TaskTracker better.
Stop re-explaining your project. Start preserving context.
β Star us on GitHub | π¦ Install from npm | π Read the docs