Free & Open Source Media Toolkit
Download media, process audio, and separate stems with AI - all running locally on your computer.
- Download from YouTube, TikTok, and more
- Auto-detect available quality (resolution, FPS)
- Batch download & playlist import
- Audio extraction (MP3, FLAC, WAV)
- Pitch shifting (-12 to +12 semitones)
- BPM detection
- Key detection
- Quick presets
- AI-powered stem separation (2/4/6 stems)
- Multiple models: Demucs, MDX-Net, VR Arch
- Output modes: All, Vocals Only, Instrumental, Drums, Bass
- GPU acceleration (CUDA)
- Built-in stem mixer with seekbar
- Node.js 18+
- Rust toolchain
- Python 3.11 (for stem separation)
yt-dlpon PATH- FFmpeg on PATH
npm install
npm run tauri:devnpm run tauri:build| Tool | Purpose |
|---|---|
| yt-dlp | Media downloading |
| FFmpeg | Audio/video processing |
| Python 3.11 | AI model inference |
| audio-separator | Stem separation |
| PyTorch CUDA | GPU acceleration |
MediaFlow runs 100% locally. We don't collect any data. See PRIVACY.md for details.
MIT License - Free to use, modify, and distribute. See LICENSE.
| Technology | Purpose |
|---|---|
| Tauri 2 | Desktop framework |
| React 19 | UI framework |
| yt-dlp | Media downloading |
| FFmpeg | Audio/video processing |
| audio-separator | AI stem separation |
| Demucs | AI model (Meta) |
| MDX-Net | AI model |
| PyTorch | GPU acceleration |
| Tailwind CSS | Styling |
| Lucide Icons | Icons |
MAMIPOKO.DEV
Special thanks to all open source contributors