AI-powered source separation (Demucs) pulls vocals, drums, bass, and melodic elements out of any track — build acapellas, instrumentals, and mashup-ready stems without a DAW.
Live demo — mute or solo a stem
Add a track to your library, or use one you've already analyzed.
One click sends the track through Demucs-based stem separation — typically done in under a minute.
Mute/solo each stem to check quality, then download the ones you need as individual files.
Creating an account and browsing your library is free, and local track analysis (BPM, key, energy, phrase detection) is unlimited on every plan. Stem separation itself is a credited feature that starts on the Hobi plan — 60 separations per month from $4.99/mo.
Four stems per track: vocals, drums, bass, and "melodic" (everything else — synths, guitars, pads). This covers the vast majority of acapella, instrumental, and mashup use cases.
Most tracks finish in well under a minute. Processing runs as a background job, so you can keep working elsewhere in the app while it finishes.
You retain whatever rights you already had to the source audio — separation doesn't grant new rights to copyrighted material you don't own or have permission to use. See our Terms of Service for details.
Yes — the underlying model isn't genre-restricted, though separation quality is generally strongest on tracks with clearly distinguishable vocal, drum, and bass elements, which covers most electronic and dance music.