I created an automation that at the very beginning I would create a transcription of an audio file in Polish received from Telegram in ogg format. I wanted to use Whisper for this, but it turns out that for some reason it only supports English.
On the other hand, when I want to select the o4 model to create a transcription, then the ogg format is not supported and I have no way to easily create such a transcription.
For what reason can’t a language other than English be selected in Whisper? After all, Whisper supports other languages. Is there an explanation for this?
Does anyone have an idea how I can solve this problem differently?