Also check out the examples: http://lav.io/2015/02/audiogrep-automatic-audio-supercuts/

And this related project: https://github.com/antiboredom/videogrep

He already states the obvious idea to integrate audiogrep into videogrep, which at the moment just uses subtitle files.

> All the instances of the phrase "time" in the movie "In Time": https://www.youtube.com/watch?v=PQMzOUeprlk

> All the one to two second silences in "Total Recall": https://www.youtube.com/watch?v=qEtEbXVbYJQ

> The President's former press secretary telling us what he can tell us: https://www.youtube.com/watch?v=D7pymdCU5NQ