I just got my code working to do the same. It’s a bit of fun converting MP3s to the correct format.

The main problem I had is that a lot of podcasts don’t just contain voices, but quite elaborate editing of other content, such as clips from TV news reports or adverts.

But if Google starts doing it it’s going to save me 26p in API charges and they will be on the receiving end of complaints.

But, it is about 99% accurate with translations.