Isovocal
Ride the vocal of any track, live.
A Max for Live device that isolates the vocals of whatever clip is playing and hands you one MIDI-mappable dial: clean instrumental on the left, isolated acapella on the right, the untouched song in the middle. In perfect sync. With zero CPU cost on stage.
Demo video coming soon
Prep at home. Perform anywhere.
Drop it on a track
Isovocal sits as an audio effect on any track with warped Session clips.
Process once
One click per song. A few minutes of offline AI separation, with progress and time remaining shown on the device.
Ride the dial
Instrumental, full song or isolated vocal — blended anywhere in between, in perfect time, at matched level.
Locked to the warp
Isovocal reads your clip's warp markers, tempo and transpose through the Live API and renders stems that stay beat-locked to Live's playback — loops, retriggers and tempo changes included.
Zero CPU on stage
Separation and alignment happen once, when you prep. During the set, playback is a buffer read. No AI running live, nothing to glitch mid-drop.
One dial, MIDI-mappable
0% clean instrumental, 50% the untouched song, 100% isolated vocal — one morph dial you can map to any knob and ride like a filter.
Real AI separation
Built on the hybrid transformer Demucs model — the open-source reference for vocal separation. Not a phase trick, not an EQ notch.
Non-destructive
No bounced stems, no extra tracks, no committing. Your clip stays exactly as it is; the device handles everything beside it.
Set-and-forget cache
Every processed song is cached. Prep your crate once and every future session with those clips is instant.
Ableton separates stems.
Isovocal makes them performable.
Live 12 can extract stems into new clips — great in the studio, clumsy mid-set. Isovocal keeps your set exactly as it is and turns isolation into a single expressive control: fade an acapella over your next track, drop to the instrumental for a vocal of your own, snap back to the record. It's a performance instrument, not a utility.