[roberdam]

that's a fantastic use case!, the easiest way (and the one i'm currently using) is by upload the audio manually on :

https://replicate.com/openai/whisper