rin.kagamine
23672387e5
The Lyria API returns a candidates[].content.parts[] array where: - parts[0] = text metadata (e.g. '<instrumental>') - parts[1] = inlineData with the actual audio payload Previous code only checked parts[0] for inlineData, which always failed with 'No audio data in response'. Fixed by iterating all parts to find the inlineData entry. Successfully generated first real audio clip (lo-fi hip hop, 30s, 728KB MP3). Also includes: .env.example, requirements updates, docker-compose fixes, frontend vite config, logger setup.
Lyria Studio
A webapp for generating, playing, and storing music clips using Google's Lyria music generation models.
Built by Kagamine Rin & Ryan.
Quick Start
# Backend
cd backend
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cd src
uvicorn main:app --reload
# Frontend (new terminal)
cd frontend
npm install
npm run dev
Architecture
- Backend: FastAPI (Python) — async Lyria API client, SQLite metadata, file storage
- Frontend: React + Vite + TypeScript + Tailwind — dark theme, audio player
- Storage: SQLite for clip metadata, local filesystem for MP3 files
API
| Endpoint | Method | Description |
|---|---|---|
/api/health |
GET | Health check |
/api/generate |
POST | Start clip generation |
/api/clips |
GET | List all clips |
/api/clips/{id} |
GET | Get clip metadata |
/api/clips/{id}/audio |
GET | Stream MP3 |
/api/clips/{id} |
DELETE | Delete clip |
Models
- v0.1: Lyria 3 Clip Preview (30s, REST)
- v0.2: Lyria 3 Pro (structure-aware, ~3min)
- v0.3: Lyria RealTime (WebSocket streaming)
License
MIT — Hardcore will never die.