Ultimate Vocal Remover
Desktop app that splits songs into vocals and instrumentals using deep neural networks.Anjok07/ultimatevocalremovergui
- Stars
- 26.9k
- Forks
- 2.1k
- Licence
- MIT
- Language
- Python
- Last update
- 30 days ago
- Type
- App
What it is
Ultimate Vocal Remover is a graphical app that uses trained source separation models to take a song and remove the vocals, leaving an instrumental track, or split out the vocals on their own. People use it to make karaoke tracks, practice with backing music, or isolate parts of a recording. The README says the core developers trained the bundled models, except the Demucs v3 and v4 4-stem models.
- Remove vocals from a song to create a karaoke version
- Separate vocals from a track for remixing or study
- Use the Time Stretch or Change Pitch tool, which the README says needs Rubber Band installed
Great for
- Musicians who want karaoke or practice backing tracks from existing songs
- Hobbyists who want to isolate vocals or instruments without using the command line
- Windows, macOS, or Linux users who prefer a point-and-click program
Not ideal for
- People who need a hosted web service with no install, since the README centers on downloadable installers and manual setups
- Users on very old or budget hardware, because the README says functionality on older or budget Macs and Windows 7 or lower is not guaranteed
- Developers who want a programmatic library, since the README describes a GUI application rather than an API
How to get it
- 1
Download the app
Clone the repository (manual install route)
git clone https://github.com/Anjok07/ultimatevocalremovergui - 2
Run it
- Download the installer for your system from the v5.6 release links in the README
- Run the installer; on Windows, install to the C:\ drive as the README advises
- Open the Ultimate Vocal Remover application once installation finishes
- Choose an audio file and a model in the interface, then start processing
- 3
Get the most out of it
- Install FFmpeg and Rubber Band if you need the Time Stretch or Change Pitch tool, following the README steps
- Use the DirectML version on Windows if you have an AMD Radeon or Intel Arc graphics card, as the README suggests
- On Nvidia GPUs in the manual install, the README recommends a CUDA build of PyTorch for faster processing
- Check the linked development update issue in the README for the latest project news
Downloads
Strengths
- Ships as ready-to-run installers that bundle Python and PyTorch, according to the README
- Offers a graphical interface instead of requiring command-line work for normal use
- Released under the MIT license and has a large community of users on GitHub
Trade-offs
- Setup can be fiddly: the README asks for extra FFmpeg and Rubber Band downloads for some features, and warns against installing on a secondary Windows drive
- macOS users may need Terminal commands to get past security prompts, and the README suggests re-enabling that security setting afterward
- Some hardware support is described as not guaranteed, so results depend on your computer
How it compares
| Project | Stars | Type | Level | Updated | Best for |
|---|---|---|---|---|---|
| ultimatevocalremovergui (this) | 26.9k | App | beginner | 30 days ago | Musicians who want karaoke or practice backing tracks from existing songs |
| demucs | 10.4k | AI model / agent | beginner | 2.5 years ago | Musicians who want to isolate vocals or drums from a song for practice or remixing |
| spleeter | 28.5k | Library | intermediate | 9 days ago | Musicians or producers who want quick vocal or accompaniment separation from a song |
| FunASR | 20.9k | Library | intermediate | 4 days ago | Developers who want to transcribe audio in Python with a pretrained model |
| koel | 17.3k | App | intermediate | 4 days ago | People who want to listen to their own music library from a browser |
| AudioKit | 11.5k | Library | intermediate | 3 months ago | Swift developers adding synthesizers, effects, or audio analysis to an iOS or macOS app |
Stats synced from GitHub. The assessment above was drafted with AI from the project's README on 2026-10-29 and may miss recent changes — check the README before relying on a command. MIT licence: free to use, modify and distribute, including commercially, as long as you keep the copyright notice.