A weekly track-curation tool for a small circle of DJs.
CLOUDER pulls fresh weekly releases from Beatport into your personal canonical library, lets you triage them with one keystroke, and ships the keepers straight into Spotify-ready playlists you can play in the browser.
Who it's for. DJs who buy or audition new releases every week and need a fast, repeatable workflow from "what came out this week" to "what's going into the set."
- Weekly automated ingest from Beatport into a personal canonical catalogue.
- Tap-to-curate workflow — one key per destination playlist, optimistic shrinks the queue.
- In-browser playback via the Spotify Web Playback SDK, keyboard-first hotkeys.
- AI-assisted screening — labels and artists are checked for AI-generated content and flagged.
- Per-DJ playlists and tags layered on a shared canonical catalogue.
CLOUDER is a serverless ingest pipeline (Lambda + S3 + SQS + Aurora PostgreSQL) plus a React 19 SPA. The backend is Python; the frontend is Mantine 9 + react-router 7. Infrastructure is Terraform.
Start here:
- System overview —
docs/architecture.md - Architecture decisions —
docs/adr/
By role:
- Backend / API / worker dev —
docs/backend/ - Data engineer —
docs/data/ - Frontend dev —
docs/frontend/ - Ops / SRE —
docs/ops/ - API consumer —
docs/api/
Local quickstart:
python -m pip install -r requirements-dev.txt
pytest -qFor the SPA: see frontend/README.md.
For deployment: see docs/ops/deploy.md.
Private — internal use only.