HU
huchukato/stemify-audio-splitter
๐ต AI-powered audio separation tool - Split any audio file into vocals, drums, bass, and other instruments using advanced machine learning. Built with React, Flask, and Facebook's Demucs v4 model.
Stemify - The Audio Splitter
A web application for audio track separation. This application allows you to upload an audio file and separate it into its components: vocals, drums, bass, and other instruments.
๐ Features
- Intuitive web interface
- Real-time track separation using AI
- Audio preview of separated tracks
- Download individual tracks
- Support for various audio formats
- Fast processing with Demucs v4
๐ Prerequisites
- Python 3.12 or higher
- uv (Python package manager) - install it with:
curl -LsSf https://astral.sh/uv/install.sh | sh - npm (Node Package Manager)
- Git
๐ป Installation
Quick Installation for Linux and macOS
- Clone the repository:
git clone https://github.com/huchukato/stemify-audio-splitter.git
cd stemify-audio-splitter- Run the installation script:
chmod +x linux_mac_install_run.sh
./linux_mac_install_run.sh- Open your browser and navigate to
http://localhost:5173
Manual Installation
- Backend Setup:
cd demucs-backend
uv sync
mkdir -p temp separated
uv run python -m gunicorn --bind 0.0.0.0:5001 --workers 1 --timeout 120 app:app- Frontend Setup (in another terminal):
cd demucs-gui
npm install
npm run dev- Open your browser and navigate to
http://localhost:5173
๐ต How to Use
- Upload an audio file (MP3, WAV, FLAC, etc.)
- Wait for the AI to process and separate the tracks
- Preview each separated track
- Download the individual tracks you want
๐ ๏ธ Technology Stack
- Backend: Flask + Demucs v4
- Frontend: React + Vite + Tailwind CSS
- Audio Processing: Facebook's Demucs v4 model
- Package Management: uv (Python), npm (Node.js)
Windows Installation
- Clone the repository:
git clone https://github.com/huchukato/stemify-audio-splitter.git
cd stemify-audio-splitter- Run the installation script:
.\win_install_run.bat- Open your browser and navigate to
http://localhost:5173
๐ต How to Use
- Upload an audio file (MP3, WAV, FLAC, etc.)
- Wait for the AI to process and separate the tracks
- Preview each separated track
- Download the individual tracks you want
๐ง Troubleshooting
- Backend not starting: Make sure port 5001 is not in use
- Frontend not loading: Check that npm installed successfully
- Audio processing fails: Ensure the audio file is under 20MB
๏ฟฝ๏ธ Security
This project maintains a strong security posture with regular vulnerability assessments and prompt patching.
Recent Security Fixes:
- โ
March 2026: Complete security audit resolution
- Fixed all 15+ vulnerabilities via npm audit
- Updated rollup (CVE-2026-27606), tar (CVE-2026-26960), minimatch (CVE-2026-27903)
- Resolved all HIGH and MODERATE severity alerts
- Status: 0 vulnerabilities (npm audit)
Security Practices:
- ๐ Regular Audits: Automated Dependabot monitoring
- ๐ Prompt Updates: Security patches applied immediately
- ๐ Vulnerability Disclosure: Transparent security reporting
- ๐ Secure Dependencies: All packages kept up-to-date
๏ฟฝ๐ License
This project is open source and available under the MIT License.
๐ฅ Author
- huchukato
- ๐ GitHub
- ๐ฆ X (Twitter)
- ๐จ Civitai - Check out my AI art models!
๐ Acknowledgments
- Facebook Research for the Demucs v4 model
- The open-source community for the amazing tools and libraries used in this project
On this page
Languages
TypeScript35.0%Python16.7%CSS12.0%Shell11.6%JavaScript10.8%HTML9.4%Batchfile4.5%
Contributors
MIT License
Created November 7, 2024
Updated March 12, 2026
