macOS activity tracker using KnowledgeC database. Track focus time, context switches, and usage patterns.
- Focus Time Analysis - Your uninterrupted usage sessions per app
- Context Switch Tracking - Identification of frequent app switching patterns
- Daily/Weekly Metrics - Usage breakdowns and trends
- Web Dashboard - Local-first visualization
- MCP Server - Query stats via Claude
Daily metrics, insights, and top focus apps with interruption patterns
Detailed context switch timeline with timestamp and app transitions
Hourly context switching trends and visual timeline of app usage
Grant Full Disk Access to your Terminal application:
- Open System Settings → Privacy & Security → Full Disk Access
- Click + and add your terminal app (Terminal.app, iTerm, etc.)
- Restart your terminal
# Install dependencies
uv sync
# Run dashboard (production mode - uses real KnowledgeC database)
SENSE_ENV=prod uv run uvicorn sense.api.main:app
# Open browser
open http://localhost:8000# Get daily stats
curl http://localhost:8000/api/stats/daily?date=2024-10-31
# Get weekly summary
curl http://localhost:8000/api/stats/weekly
# Top context switches
curl http://localhost:8000/api/switches/top?limit=10- Backend: FastAPI + SQLite (KnowledgeC)
- Cache: In-memory (dict with TTL)
- Frontend: Jinja2 templates
- Packaging: uv + pyproject.toml
- All data stays local
- Read-only access to KnowledgeC
- No external services
# Install dev dependencies
uv sync --all-extras
# Sync the KnowledgeC database (for development mode)
./scripts/sync_db.sh
# Run tests
uv run pytest
# Run locally (dev environment - uses data/knowledgeC.db)
uv run uvicorn sense.api.main:app --reload
# Run with production database (uses ~/Library/Application Support/Knowledge/knowledgeC.db)
SENSE_ENV=prod uv run uvicorn sense.api.main:app --reload- macOS with Full Disk Access
- Python 3.11+
- Access to
~/Library/Application Support/Knowledge/knowledgeC.db
If you find Sense useful:
- ⭐ Star this repo to show your support
- ☕ Buy me a coffee to fuel development
Your support helps keep this project maintained and improved!
MIT