This is the Python microservice for audio stem separation using Demucs AI.
- Receives audio file URLs via REST API
- Uses Demucs 4.0.1 to separate audio into stems (vocals, drums, bass, guitar, other)
- Returns base64-encoded stem audio files
POST /separate- Separate audio into stemsGET /health- Health check
This service is designed to be deployed to Railway as a separate service from the main Node.js frontend.
PORT- Port to run the service on (automatically set by Railway)
pip install -r requirements.txt
python demucs-api.py