Introducing our Audio-to-MIDI model
Turn any sound into editable MIDI and MusicXML.
Have you ever listened to a song you loved and wished you could make it your own? Developed by Mirelo and kyutai, our audio-to-MIDI model lets you do exactly that. Import a song and it generates the MIDI for you, separated into each instrument.
Features
Our open-source model is built for the full mix use case. Upload a recording and it gives you:
Automatic instrument detection.
The model finds each part in the mix and labels it (voice, drums, electric bass, synth lead, and so on).
A separate MIDI track per instrument.
Every detected part comes back as its own track on a piano roll, where you can review the transcription and adjust any notes you want to change before you export.
Musical context.
Leverage our model's understanding of isolated notes to build applications like Sound-to-MIDI that delivers the full harmonic picture and detect chords, key, and tempo.
One-click export.
Solo, mute, or download any track as MIDI and drop it into your DAW.
Built on research we're sharing openly
Audio-to-MIDI runs on a model we developed with kyutai, and we've open-sourced it for anyone who wants to build with it. The version in Mirelo Studio is our own, trained on a larger dataset for even better results.
Pick a song. Hear it come apart.
Audio-to-MIDI is free to try in Mirelo Studio — no setup, no software to install.
