JeffreyCA/spleeter-web
Self-hostable web app for isolating the vocal, accompaniment, bass, and drums of any song. Supports Spleeter, Demucs, BS-RoFormer. Started in 2019.
What it solves
Spleeter Web provides a user-friendly web interface for audio source separation, allowing users to isolate or remove specific components of a song, such as vocals, accompaniment, bass, and drums, to create instrumentals or vocal-only tracks.
How it works
The application integrates several AI-driven audio separation models, including Spleeter, Demucs (v3 and v4), and BS-RoFormer. It uses a Django backend API, a React frontend, and Celery for managing the memory-intensive task queue of audio processing. Users can import audio via file upload or YouTube links, and the system supports both CPU and NVIDIA GPU acceleration for faster processing.
Who it’s for
It is designed for musicians, producers, and audio engineers who need to separate audio stems from existing tracks without needing to run complex command-line tools.
Highlights
- Dynamic Mixes: Real-time playback and export of custom mixes of separated components.
- Multi-Model Support: Access to Spleeter, Demucs, and BS-RoFormer models.
- Flexible Import: Support for direct file uploads and YouTube integration via API.
- Infrastructure Ready: Fully Dockerized with support for GPU acceleration and cloud storage backends like AWS S3 and Azure Blob Storage.
相關
- 專案
- 專案
- 專案
- 專案
- 專案