Scan a paper calendar (school, lunch menu, sports, activities), review the parsed events, and export an .ics file or a subscribed feed for Apple and Google Calendar.

Self-hosted only: runs directly with Python or as a Docker image, with SQLite as the backend. The page is read by a vision model on your own Ollama server (or, optionally, Claude).

Product & technical spec

Status: Phase 1 (MVP for one household) in progress.

Run it

Copy .env.example to .env and set at least PUBLIC_URL and OLLAMA_URL.

Plain Python (Python 3.11+, and Node only to build the front end):

cd frontend && npm ci && npm run build && cd ..
pip install -r requirements.txt
python main.py

Then open http://127.0.0.1:8000. Rebuild the front end after changing anything in frontend/.

Docker: copy compose.example.yml to compose.yml, then docker compose up -d --build.

Backups: python main.py backup <folder> copies the database (safely, while running) and the photos. Schedule it nightly.

Develop

cd frontend && npm run dev serves the front end with hot reload and proxies /api to python main.py.
Tests: python tests/test_phase1.py, python -m parentcalendar.ics, python -m parentcalendar.extraction --selftest.
Extraction quality (Phase 0): python phase0/extract.py, then python phase0/score.py private/runs/<run>. The golden set lives in private/, outside git.