The application is a little like Audacity, though it has nowhere near as many features, and a lot more bugs. I originally started programming it as a basis for my own text-to-speech research. You can load a .WAV file and play it back. You can select parts of the sound, give those selections labels, and then play back your labeled selections in some new order. You can zoom in and zoom out. A few rudimentary filters are provided that allow you to silence or amplify selections, or generate a sine wave of a given frequency. You can even export your labeled selections as an subtitle file in .SRT format.
One notable missing feature that I'd eventually like to have is the ability to record from a microphone, but that's somewhat tricky and browser-dependent at the moment. I'd also like to be able to mix all the tracks in the session into a single .WAV, and of course just improve the experience overall.
If you're interested, you can test the application by using it to load a pre-prepared session file that will hopefully be made available at http://thiscouldbebetter.neocities.org/JFK-We_Choose.json. Click the "Load Session" button, select the session file in the load dialog, then click the "Play Tagged Selections" button to hear an amateurishly remixed version of John F. Kennedy's "We Choose to Go to the Moon" speech, in which he says the opposite of that.
It's good clean fun!