Best Approach For Analyzing Imagery (Written Paper)

API
, , , ,
1

Hi there,

I have an AI Feedback Generator that I would like to provide feedback on image uploads (written text).

What is the best approach + endpoint for this use case?

Should I extract the text from the image, then send the text to the Chat Completions endpoint? Or use the Assistant endpoint to achieve this?

Thank you for your recommendations.

2

Welcome to the community!

You might try OCR first then send the text to Chat Completion. The vision models are getting a lot better, though, so one of the newer ones might work.

Have you tried anything yet?

If you search the forums, you should find some relevant information.