Backend (FastAPI): - User auth with email/password and Spotify OAuth - Spotify playlist import with audio feature extraction - AI recommendation engine using Claude API with taste profiling - Save/bookmark recommendations - Rate limiting for free tier (10 recs/day, 1 playlist) - PostgreSQL models with Alembic migrations - Redis-ready configuration Frontend (React 19 + TypeScript + Vite + Tailwind): - Landing page, auth flows (email + Spotify OAuth) - Dashboard with stats and quick discover - Playlist management and import from Spotify - Discover page with custom query support - Recommendation cards with explanations and save toggle - Taste profile visualization - Responsive layout with mobile navigation - PWA-ready configuration Infrastructure: - Docker Compose with PostgreSQL, Redis, backend, frontend - Environment-based configuration
1.2 KiB
1.2 KiB
Vynl - AI Music Discovery App
Project Status: Ready to build MVP
Design Doc
See DESIGN.md for full product spec.
Quick Summary
- Import playlists from Spotify/Apple Music/YouTube Music/Tidal
- AI finds new music you'll love and explains why
- Free tier (10 recs/day, 1 playlist) + Pro ($4.99/month, unlimited)
- Cross-platform import/export
Tech Stack
- Backend: Python 3.12+ / FastAPI / PostgreSQL / Redis / Celery
- Frontend: React 19 + TypeScript + Vite + Tailwind CSS
- AI: Claude API for recommendations + Spotify audio features + MusicBrainz + Last.fm
- Auth: Email + OAuth (Spotify, Apple, Google)
- Payments: Stripe
- Mobile: PWA (progressive web app)
MVP Scope (v1.0)
- User auth (email + Spotify OAuth)
- Spotify playlist import
- Manual song/artist search
- AI recommendations via Claude API
- "Why you'll like this" explanations
- Taste profile display
- Save/bookmark recommendations
- Responsive web app (PWA)
Brand
- Colors: Deep purple (#7C3AED) + warm cream (#FFF7ED) + charcoal (#1C1917)
- Vibe: Warm analog nostalgia meets modern AI
Git
- Repo: To be created on Gitea (chrisryn/vynl)
- Branches: master + dev (standard workflow)