Drag a note, draw a curve, swap an instrument — the document rewrites itself underneath, and you never have to look at it. Or open it and edit the line directly. Both are the same edit, in the same history, with the same undo. What you touch locks; the studio's agents keep working around it. How the language works →
"Longer intro, later drop, darker middle eight." The arranger recomposes sections, energy arcs and transitions to your brief — at the next musical boundary, never mid-phrase.
Describe a sound — "hollow woody knock, metallic tail" — and it designs the instrument as a validated patch graph, or renders a sample kit from a recipe, mid-session, in under a second.
Gain staging, spectral balance and dynamics are scored against profiled reference records — the same measurements a mastering engineer reaches for, continuously evaluated.
Your bassline stays your bassline while the studio keeps working around it. Every change — yours or an agent's — is an invertible diff in the session history. Undo is exact.
The tracks you drag and the code streaming past are the same thing seen twice — real values from the engine, bar by bar, not a visualization of one. Move a note and the line changes; change the line and the note moves. Tap, edit, hear it next bar. It is also the same language the agents write, so directing them and editing yourself are one continuous skill — and watching it is how you learn to read it.
Every session is readable JSON — save it, share it, re-import it, replay it bit-identically. The session is the product: a portable program the engine performs.
Velocity-precise step editing, the full instrument rack with exact machine parameterizations, sends and automation lanes. On iPad the Pencil writes velocity with pressure; on the Mac the full keyboard drives it.
Bounce per-track stems from the deterministic renderer, and take the full session as JSON. MIDI and automation export are on the roadmap. Kontinuum is where tracks are born — finish and master them wherever you already work.
Point the studio at a track in your own library. It analyzes structure, harmony, groove and per-stem character, picks or designs the instruments — and hands you a new editable session derived from its musical characteristics, correctable line by line. Analysis only; nothing is copied, and the audio never leaves your device.
Package the way you think — groove, harmony habits, rack, sample recipes, mix character — as a Soul. Publish it free, or license it: 80% of every use goes to you, metered by real listening, not playlist politics.
The app installs with Listen; the editing surface appears the moment you want it — no extra purchase, no separate app, same document underneath.