openai-audio-create-transcription doesn't work

[agzam]

Tried a few different ways but couldn't make it work. Maybe openai has changed something about the endpoint? I keep gettin invalid_request_error

[jcs090218]

Can you post your code here? That way, it's easier for me to debug! Thanks!

[agzam]

Even something as basic as:

(openai-audio-create-transcription
 "my-file.mp3"
 (lambda (data)
   (print data)))

[jcs090218]

Just to let you know, I'm unable to get this to work as well. I would need to investigate further more. 🤔

Sorry for the time you have wasted! 😓

[agzam]

Sorry for the time you have wasted!

Hey, not at all. As an avid Emacs user I am in fact very excited to find this project and happy to know that someone is working in that space. My use case for transcription feature is very simple and straightforward. I'm hoping to write a simple set of functions that would allow transcribing any .mp3, .mp4 directly from Dired. Of course, the tokenization and chunking up large files would be a challenge, but it would be great to make something like that using only Emacs Lisp, right?

[jcs090218]

I have done some research, and it seems like many people are facing the same issue, and it remains unresolved! Some links below:

• https://community.openai.com/t/400-bad-request-error-when-passing-audio-to-server-before-passing-to-openai/100908
• https://community.openai.com/t/calling-whisper-api-using-curl-request-keeps-giving-error/81510

But to be honest, I don't know if I made the right implementation or not. And indeed, they have changed their API endpoint, which I need to update as well. 🤔