Changelog
Yuma can now write automation in Ableton
This is Yuma's deepest Ableton release so far. The headline is automation: Yuma can shape a sound over time now, not just place notes. It also got a serious reliability pass so the app starts cleanly on every machine.
- Yuma writes filter sweeps, volume rides, panning, and other parameter automation into your clips.
- The app repairs its own saved data on launch and falls back to a recovery screen instead of a blank one.
- A single universal Mac app runs natively on Apple Silicon and Intel, with automatic updates.
- Improved MIDI chord generation that no longer repeats the same voicings on every request, so re-rolling gives you new ideas.
- Vocal mode stays consistent when you ask Yuma to revise an acapella it already made.
Generated audio is now real, findable files
Audio you make in Yuma stopped living inside the app and became real files you can find, drag, and keep. You can also talk to Yuma now instead of typing every request. Drum kit export was rebuilt to work the same on any machine.
- Audio saves as standard WAV files in your Documents/Yuma folder, not hidden in app storage.
- Every audio, MIDI, and kit card has a button that opens the exact file in Finder.
- Realtime voice dictation lets you speak your requests out loud.
- Drum kit export works across macOS, Windows, and multiple Ableton versions, and falls back to plain samples when a rack can't be built.
- Asset cards recover their file and state instead of breaking if a generation gets interrupted.
Stems, new audio engines, and Serum presets
This update widened what Yuma can actually make. On top of MIDI and vocals, it can pull a finished track apart, reach for new generation engines, and design synth patches. The workspace got a cleanup at the same time.
- Split any track into separate vocal, drum, bass, and instrument stems.
- New generation engines cover a wider range of sounds and styles.
- Generate custom Serum patches and load them straight into the synth.
- A reworked empty state and sidebar, with starter prompts that wrap into rows instead of hiding behind a More menu.
Custom drum kits and dictation
Yuma moved past single sounds and started building whole kits. It also gained its first voice input, so you could talk to it instead of typing. Every generated file moved onto one consistent card.
- Ask for a kit and Yuma builds a full folder of samples plus an Ableton Drum Rack you can drag straight in.
- Dictation lets you speak your prompts instead of typing them.
- MIDI, audio, and kits share one card now, with preview, download, and reveal in the same place.
AbletonGPT is now Yuma
AbletonGPT became Yuma on April 30, 2026. It is the same product, the same team, and the same account, so you can sign in with the email and password you already used. The relaunch was a major upgrade, and every early supporter got 4,000 bonus credits added as a thank-you.
- Download the new desktop app for macOS or Windows from the pricing page.
- A better MIDI engine, with cleaner chord voicings, on-grid timing, in-key approach chords, and arpeggios that actually sound musical.
- Arrangement mode writes MIDI directly into Ableton's Arrangement view, not just Session clips.
- Thinking mode is on by default for the models that support it, so the agent plans before it acts.
- Audio generation for vocal acapellas, instrumentals, and sound effects from a text prompt.
- Every plan now includes a monthly credit balance, with no external Anthropic or OpenAI key to bring.
- Generated MIDI and audio drag straight into Logic, FL Studio, Pro Tools, or any other DAW.