No unfortunately not. Maybe I'll put it on my GitHub soon. I was thinking maybe I'd find some way to make money on this, so I haven't released the source yet.
As to how it works, internally there's a modular data processing pipeline. I wrote a number of modules that serve a specific purpose (sine wave, square, etc, and reverb, gain, etc). They all get connected together in a driver class. Each driver makes heavy use of random numbers. Those in turn decide the notes played and durations. The durations are extremely short, leading to a granular synthesis of sorts. The notes are chosen from classes I structured around standard chord theory and classical modes.
I have an unreleased version that processes mic input, but it's not ready for prime time yet.
As to how it works, internally there's a modular data processing pipeline. I wrote a number of modules that serve a specific purpose (sine wave, square, etc, and reverb, gain, etc). They all get connected together in a driver class. Each driver makes heavy use of random numbers. Those in turn decide the notes played and durations. The durations are extremely short, leading to a granular synthesis of sorts. The notes are chosen from classes I structured around standard chord theory and classical modes.
I have an unreleased version that processes mic input, but it's not ready for prime time yet.
Edit: I decided to release it, probably under AGPL license. Keep an eye on this page: https://github.com/hollingsworthd/ApplPi