A web application that separates audio into stems (vocals, drums, bass, guitar, other) using Demucs. Supports uploads, YouTube URLs, and Spotify links.

A preview of the interface

• Upload audio files for stem separation
• Extract stems from YouTube videos
• Extract stems from Spotify tracks
• High-quality separation using Demucs' htdemucs_6s model
• Easy-to-use web interface built with Gradio

The most convenient way to run the application is using Docker. Two commands and you're done!

1. Get the latest Docker image:

   docker pull darioloi/stemsep

2. Run the container:

   docker run -p 7860:7860 darioloi/stemsep

   This will run the app without GPU support.

3. If you want to use GPU acceleration, ensure you have NVIDIA Container Toolkit installed. Then run:

    docker run --gpus all -p 7860:7860 darioloi/stemsep

4. Access the web interface at http://localhost:7860

If you prefer to run the application locally, you can install it using pip. It's recommended to use a virtual environment to avoid conflicts with other packages. Also check out uv for a better local python experience.

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

device = torch.device("cpu")

1. Clone the repository:

   git clone https://github.com/dario-loi/stemsep.git
   cd stemsep

2. Install dependencies:

   pip install -r requirements.txt

3. Install ffmpeg (required for audio processing):

   • Ubuntu/Debian: sudo apt-get install ffmpeg
   • macOS: brew install ffmpeg
   • Windows: Download from ffmpeg.org

4. Run the application:

   python app.py

5. Access the web interface at http://localhost:7860

1. Run with Docker Compose, this will automatically enable GPU support

   docker-compose up --build

2. If you want to run it without GPU, use:

   docker-compose -f docker-compose.cpu.yml up --build

3. Access the web interface at http://localhost:7860

1. Visit Spotify Developer Dashboard
2. Create a new application
3. Set environment variables:

   export SPOTIFY_CLIENT_ID="your_client_id"
   export SPOTIFY_CLIENT_SECRET="your_client_secret"

For Docker, add these variables to your environment:

docker run -p 7860:7860 \
  -e SPOTIFY_CLIENT_ID="your_client_id" \
  -e SPOTIFY_CLIENT_SECRET="your_client_secret" \
  audio-stem-separator

For Docker Compose, add to docker-compose.yml:

services:
  stem-separator:
    environment:
      - SPOTIFY_CLIENT_ID=your_client_id
      - SPOTIFY_CLIENT_SECRET=your_client_secret

1. Access the web interface at http://localhost:7860
2. Choose a tab based on your input (Upload, YouTube, or Spotify)
3. Provide the audio/URL and click "Process"
4. Download the separated stems