It’s open source so others might expand on it some day. Basically you let it analyze an audio recording you make and it will spit out a MIDI file, and you have parameters to control to adjust how good/accurate or whatever the MIDI is.
Then you put that MIDI file into a DAW, put a synth on that track (or maybe a guitar sampler? Whatever you want really)
Demo from Bad Snacks using a harp, voice, and violin: