This fork adds support to multiple audio files conversion.
See also: A local server piano-scribe (Website)
This app converts raw audio to MIDI using Onsets and Frames, a neural network trained for polyphonic piano transcription.
Record yourself playing piano or choose audio files with solo piano from your device to transcribe!
Don't have a piano? Try singing to it to see what your voice sounds like played by a piano! Your voice isn't a piano though, so it might not sound like you at all 😅.
All of the processing happens locally in the browser using Magenta.js and TensorFlow.js.