Ultimate Vocal Remover

Ultimate Vocal Remover GUI provides a graphical interface for applying state‑of‑the‑art source‑separation models to audio files, allowing users to isolate or remove vocal tracks. The application bundles the necessary Python runtime, PyTorch, and model files, so it runs without external prerequisites beyond the operating system. It supports both Windows (10 or newer) and macOS (Big Sur or newer), with optional GPU acceleration via DirectML on AMD/Intel cards, CUDA on Nvidia, and MPS on Apple Silicon.

The software is aimed at musicians, producers, podcasters, and anyone needing to create karaoke or instrumental versions of recordings. It includes pre‑trained models, except for Demucs v3/v4 four‑stem variants, and offers additional tools such as time‑stretching and pitch shifting when the Rubber Band utility is installed. Users can install via a bundled installer or manually by extracting the repository and installing required dependencies like FFmpeg.

Stability is emphasized for supported hardware configurations, and the macOS build adds GPU‑accelerated processing for Demucs v4 and MDX‑Net models on M1 devices. The project is maintained by its core developers and is released as a stable version.