NeuralNote
Audio Plugin for Audio to MIDI transcription using deep learning.
What is NeuralNote?
Audio Plugin for Audio to MIDI transcription using deep learning.
Overview of NeuralNote
What it is
NeuralNote is an audio plugin for audio-to-MIDI transcription using deep learning. It runs fully locally on the user's machine. Version 2 uses the MuScriptor model, a transformer with 103 million to 1.4 billion parameters, replacing the previous Basic Pitch model.
Key features
- Multi-instrument transcription: transcribe full mixes with notes grouped by instrument.
- Built-in multi-instrument synth with per-instrument gain, mute, and solo.
- Fully local processing: audio never leaves the machine; only downloads models and checks for updates.
- Supports recording live audio or dropping audio files (.wav, .aiff, .flac, .mp3, .ogg, .m4a on macOS).
- Drag-and-drop MIDI export to DAW or save as MIDI file.
- Three model sizes (small, medium, large) trading speed for quality.
Use cases
- Transcribing polyphonic audio from recordings or live input into MIDI for further editing.
- Extracting individual instrument parts from mixed audio for remixing or arrangement.
- Converting vocal or instrumental melodies into MIDI for use in virtual instruments.
Getting started
Download the latest release from the website or releases page. Install the appropriate package for your OS. On first launch, download a transcription model from within the plugin. Apply NeuralNote as an AudioFX plugin on a track, record or drop audio, select instruments, and click Transcribe. Drag the resulting MIDI to a MIDI track or save it.
License and pricing
Licensed under Apache-2.0. Open source.
Based on the project's public README and repository metadata.