It's open-source so you can see how everything is done. The audio processing is all done with Web Audio APIs.