Is this something you maybe could use WebRTC for?
Making an audio-recording is probably simpler, use the Web Audio API to record and then push the file to a cloud service. Making comments is just annotating text to timestamps in the audio. Soundcloud does this in a good way.
For a truly bleeding edge solution you could use the Web Audio API, record to a buffer, and then distribute the audio using WebTorrent :)