From the first idea
to the last note.
A native music workspace with the arrangement and conversation side by side.
Simplified interface illustration. Example musical content.
Talk through
the next change.
Ask for a new part, a different rhythm, or a quieter ending. Chat can inspect the song and edit its tracks, notes, instruments, and arrangement.
Keep several conversations attached to one song. Successful changes apply as one undoable edit.
Focus on the music.
Compose through AI and inspect every part. Recording, import, transcription, export, and Undo are always available.
⌃ ⌘ 1A melody can
start with you.
Record or import a solo whistle, hum, sung melody, or beatbox take. Review detected events, choose your instruments, preview, then add to the song.
- Preserve measured timing, including rests
- Layer compatible instruments
- Add optional AI accompaniment
Detection is approximate. Singing mode detects melody, not lyrics. Preserve Time with one layer runs locally.
Find the right
sound for a part.
Browse six built-in synth and electronic sounds, plus 45 sampled sound choices across 42 downloadable FreePats banks.
Preview installed sounds explicitly. Keep downloads available offline across your songs. Missing instruments are shown clearly, with no silent substitution.
Bring a reference.
Explore its parts.
Attach audio to a conversation. Local tools can separate stems and estimate editable notes to help with recreating, restyling, or arranging a transition.
Apple Silicon and explicit local model setup required. Extraction is approximate; original vocals remain audio by default. Optional cloud audio inspection is off by default.
Save the whole song.
Share the music.
A .unableton project keeps your music, audio, chats, and generation history together. Export stereo WAV to share a mix or multitrack MIDI to continue elsewhere.
Record and import audio too. Optional score-to-singing tools use a separate provider and key.
Your next song
can start here.
Download the preview and try an idea.