I suspect that for rights reasons they basically have to be newly opened and transcribed each time. It the same with the news feed where when you want to add one of the articles, it still has to go to the website and open it for each individual reader.
Yes, that is realistically how it will work. Over time, if we have a better solution, we can look into it. Some podcasts do have transcripts already but if transcribing is required, that will be done for each user.
If you implement it that way, I’m afraid regular podcasts listeners - like me - won’t use it: I have a lifetime subscription for Spanish and I’d already burn my 600 monthly minutes just by listening to 1 - one - morning daily podcast (El País).
My podcast reader client - SnipD - also enforces a minutes quota for transcriptions, BUT it’s only billed to the first user to request the transcription. The subsequent users also don’t have to wait 20 minutes for the audio transcription step, as it’s already been done.
I’ve never had to “pay” nor wait for super popular podcasts episodes, only for niche ones.
I guess they’ve ironed out the legal aspect of sharing internally the transcription files generated by them.
Another scheme would be to charge us 1 minute - f.ex - for any episode, popular or not. That could be a win-win.