I have not had to deal with that, but I have idly thought that it might be easier to pipe the audio version into google assistant or something, and see what it comes up with.