Hi, we have been processing gpt-4o vision requests a lot yet all them cannot load in images in the request. Therefore, I'm not able to analyze if user submitted requests are accurate - any idea how to fix this? Thanks!

example of a request ID: 8b851323-33e9-4235-9e6b-8d71e510a78f

I implemented this afternoon of passing Helicone-Request-Id with a custom uuidv4 to the header to implement a thumbs up and thumbs down feedback. However, amongst implementing it, I realise responses are missing despite checking that the code works for the one API route I have (i.e it logs the response some times but misses it sometimes with the same route).

I then rolled back to previous version without the Helicone-Request-Id and the response about seems to start rolling back up.

During the time when in production I implemented setting the custom Helicone ID, I have searched for some certain responses I did myself which were missing too.

Any ideas for this? Having the user thumbs up thumbs down feedback will be nice data for us to have for finetuning.

Hi weirdly once i inject images into gpt-4o (via azure) all the token count and cost tracking displays “not found”. Normal content: prompt (string) would work. But once i declare an image type and pass an image to gpt-4o it just displays not found