Help & support.
Vocal and stem separation that never uploads a note. Here's how to get help with it.
Email a human.
direct · no ticket queueEmail us about anything — a bug, a question about a separation mode or the weight download, or a feature request. Because OhmSplit runs entirely on your Mac, please don't attach your audio to the message; we can help without it. For non-private questions you can also open an issue on GitHub.
Email support ↵The short answers.
privacy · stems · formatsNo. Separation runs 100% on your Mac — the track you drop in never leaves your machine. There's no account, no telemetry, and no analytics; the privacy manifest declares no tracking. The only network access is the first-run download of the separation model weights themselves.
Two modes: a 2-stem split into vocals and instrumental — the acapella / karaoke workflow — or a 4-stem split into vocals, drums, bass, and other. You pick the mode per track, preview the result in the synchronized multitrack mixer with per-stem solo, mute, and level, then export the stems you want.
A curated set from the Ultimate Vocal Remover model family: Facebook's HT-Demucs for 4-stem separation, and Kim Vocal 2, MDX23C InstVoc HQ, and UVR-MDX-NET Inst HQ 3 for high-quality 2-stem vocal/instrumental splits. All of them run locally on your Apple Silicon Mac.
Input is anything ffmpeg can read — MP3, WAV, FLAC, M4A, and OGG all work. Stems export as WAV by default, with MP3 and FLAC also available, downloadable per-stem or as a single zip.
Only once. On first separation, your Mac downloads the model weights directly from the public Hugging Face mirror — we never proxy or see the transfer. After the weights are cached, separation is fully offline.
OhmSplit is a one-time purchase of $4.99 on the Mac App Store. No subscription, no per-track credits, no upload quota — you buy it once and split as many tracks as you like.