Your studio.
Everywhere.

You can convert 2 projects a month with no account at all. A free account gives you 120 credits a month — one pool for every tool: a conversion is 1 credit, a 30-second generation 10. Pro is $12/mo for 2,000 credits a month, plus extra usage if you need more. The engine is identical on every plan; paying lifts the cap, never the quality. See pricing.

No. The studio is a web app — open a project in the browser and edit tracks, MIDI, the mixer and plugins, with everything saved to your account as you go. There's a desktop app if you want folder-watching and conversion that never touches the network, but nothing needs installing to start.

Logic Pro, Ableton Live, FL Studio, Pro Tools and REAPER — a free online DAW converter for every direction between the five, with the direction read from the file you drop (.logicx, .als, .flp, .ptx, .rpp): als to logicx, flp to als, ptx to logicx, rpp to als and the rest. Tracks, MIDI, audio, tempo maps and markers all arrive editable. We'd rather say it plainly: no Cubase or Studio One yet. See exactly what carries over.

Yes. Share a session as view, comment or edit, and you're both in it live — each other's cursors, who has which track selected, and edits landing as they happen. Share links last 30 days by default. If you both change the same track at the same moment the later edit wins, but you're told when it happens and the earlier version is still in the session's History.

No — 15 sources across strings, brass, winds, keys, guitars, voice and rhythm, modelled rather than recorded. Velocity moves the timbre instead of the fader, and articulations that genuinely change the sound — pizzicato, mute — are separate sources rather than effects laid on afterwards. They're what every generative tool plays.

For conversion, it doesn't have to: switch on Browser mode, pick the project folder, and your audio stays on the device — only the session file and a few kilobytes of clip headers upload, and the converted project is assembled locally. The generative tools (cover, separation, transcription) run on our GPUs, so that audio does upload; it's deleted from processing once your result comes back.

No. Your uploads, your generated audio and your sessions are never added to our training corpus. That's architectural rather than a promise in a policy — your content is stored apart from training data, where the training pipelines cannot reach it. The privacy policy says it in full.

Yes — an API and an MCP server cover separation, transcription, audio-to-session, cover, generation and project conversion, so an agent or a script can drive the same engine the site does. Mint a key in settings.