GitHub radar
Free Local Audio Stem Separation: StemDeck
Split any audio track into vocals, drums, bass, guitar, piano, and more — locally on your machine, no account, no upload, no subscription.
StemDeck is a free, open-source desktop application that splits audio into up to six stems — vocals, drums, bass, guitar, piano, and other — entirely on your machine using Meta's Demucs AI model. Drop in an MP3, WAV, FLAC, OGG, MP4, or M4A file, or paste a YouTube URL, and StemDeck handles the rest. A DAW-style waveform editor lets you mute, solo, and adjust levels for each stem, loop a region, and export individual tracks or a custom mix. On first launch the app downloads the Demucs model (~170 MB) and Python runtime, then works fully offline afterward. macOS and Windows builds are available as ready-to-run installers — no Python or system dependencies required.
Why a vibe-coder should care
No subscription, no account, no upload — your audio stays on your machine. Useful for creating karaoke tracks, practicing a specific instrument part by ear, remixing, or isolating vocals for content production.
How to install
Copy this and send it to your agent — Claude Code, Codex, any of them:
Set up StemDeck from https://github.com/stemdeckapp/stemdeck — download the installer for my platform from the Releases page, launch it, and walk me through loading an audio file and getting the separated stems. Ask if you need any clarification.
Runs on a regular laptop — a MacBook with an M-chip runs fast via GPU. Slower on Intel Mac or without an NVIDIA GPU, but still works. Around 700 MB of free space needed for the model and environment.
Open on GitHub▌ More finds