Generate an image of a madman playing drums. It should be chaotic and nihilistic.
Drummer is an app used to practice your favorite songs. Upload an MP3 file or provide a YouTube URL and it will strip drums from that song using AI-powered source separation and provide an updated MP3 without the drums.
It provides a simple, but elegant, web UI that allows for uploading files, downloading from YouTube, and managing songs, as well as a table list of all previously processed and stripped songs, with the ability to delete and rename songs.
Drummer uses Spleeter, Deezer's AI-powered source separation library, to accurately separate drums from your music. Spleeter uses deep learning to isolate different instruments and vocals, providing much higher quality drum removal compared to traditional center-channel extraction methods.
The application uses Spleeter's 5-stem model for highest fidelity separation into:
- Vocals
- Drums
- Bass
- Piano
- Other instruments
The final output combines vocals, bass, piano, and other instruments while excluding the drums, giving you a clean, high-fidelity backing track for practice. Audio is processed using the highest quality settings to preserve acoustic accuracy.
Written in Go and deployable in a Docker container. Uses Spleeter for AI-powered drum separation and FFmpeg for audio processing. Supports both file uploads and YouTube URL downloads using yt-dlp. Song metadata is persisted using SQLite database for tracking across container restarts.
Written in React and included with the deployed Docker container.
- Database: SQLite database stores song metadata (name, file paths, upload date)
- Files: Original and processed audio files are stored in mounted volumes
- Volumes: All data persists across container restarts via Docker volumes
- Docker and Docker Compose
- Clone this repository
- Run the application:
docker-compose up --build
- Access the application at
http://localhost:8080
The first run may take longer as Spleeter downloads its pre-trained models when processing the first song.
Once the application is running, you can:
- Upload MP3 files: Click the upload button to select and upload MP3 files from your computer
- Download from YouTube: Enter a YouTube URL to download and process the audio
- Manage songs: View, rename, delete, and download your processed songs from the songs table
Both uploaded files and YouTube downloads are processed through the same high-quality drum removal pipeline.
The application creates the following directories on your host machine:
./uploads/- Original MP3 files./processed/- Processed MP3 files without drums./data/- SQLite database file./temp/- Temporary files during processing
All your songs and metadata will persist across container restarts and rebuilds.