AI Mixing and Mastering From Your Stems — Full Walkthrough
The whole application end to end: stems in, a finished mix and master out, every decision written in plain language.
Nine short videos. The first covers the whole application in under three minutes; the rest take one thing each.
Nothing loads from YouTube until you press play.
The whole application end to end: stems in, a finished mix and master out, every decision written in plain language.
Two decision engines, how to switch, and what actually changes in the mix.
What the offline built-in engine measures, how to read its Engine Notes, and every direction word it understands.
Making one stem sit forward and pushing another back, without touching a fader by hand.
Moving the faders after the master is finished, hearing it loudness-matched, and saving the result as a profile.
Matching a reference, reading what it matched and by how much, and exporting stems in either of two modes.
The six built-in profiles, and saving a mix you like as your own so every future master starts from it.
Connecting the optional AI engine with your own key — provider, model, and exactly what gets sent.
Steering a mix by typing plain English, and how a follow-up builds on the last one instead of starting over.
The built-in analysis engine runs offline and is the default — no account, no key, no connection. The optional AI engine uses your own API key and receives the measurements only, never your audio. Both drive the same EQ, compressors, reverb and limiter; what changes is what gets asked for.
The demo is the complete application with exports disabled. Windows 10 and 11, and macOS 15 or newer on Apple Silicon — M1 or later. Intel Macs are not supported.