Ultimate Vocal Remover
An open-source desktop vocal remover that processes audio locally on Windows, macOS and Linux using neural source separation models.
Ultimate Vocal Remover is a desktop app for separating vocals and other stems from audio files on your own computer. It's for people making karaoke backing tracks, isolating a vocal, or working with separate parts of a song. It runs on Windows, macOS and Linux, and its MIT license allows you to use and modify the software.
The graphical interface brings neural source separation models into a desktop workflow. Supported models include MDX-Net and Demucs, including Demucs models that separate music into four stems. The core developers trained most of the models supplied with the app; the Demucs four-stem models are exceptions. That choice of models matters if you want to try different approaches to separating the same recording.
GPU support depends on your hardware. The app can use compatible NVIDIA GPUs, with a GTX 1060 with 6GB of VRAM listed as the minimum for GPU processing and at least 8GB recommended. On Apple M1 Macs, GPU acceleration supports Demucs v4 and all MDX-Net models. AMD Radeon support is limited.
The Windows bundle targets Windows 10 or later, while the Mac bundle targets macOS Big Sur or later. Bundled desktop downloads include Python, PyTorch and the app's dependencies, so you don't need a separate Python setup. WAV files work without FFmpeg; processing other audio formats requires FFmpeg.
