I'd like to add to this that the transcription doesn't even need to be 100% accurate to be useful. It'll mostly be used to get the gist of what was said and target more specifically which audio message requires to be re-listened to instead of having to listen to everything again.