I am using the API with an assistant to extract data from .pdf files. This has worked great, but it suddenly stopped working.
I suspected perhaps something was failing in my api calls, but just using the assistant in the playground gives the same error.
- gpt-4o-mini
- An easy instruction to extract the relevant data I want (This has not changed before or after the issue appeared)
- Uses file search with no other advanced instructions
- I upload a .pdf file to the message along with a simple prompt saying what data I want extracted (The .pdf files I have tested with until now didn’t work, so I tested an almost empty .pdf with only a few lines and it made no difference)
- I submit the run and it chugs for a while before returning the following error:
"code":"vector_store_timeout" "message":"Failed to process files in vector store vs_67a626eed1..... Please ensure all files in the vector store are processed before retrying the run."
Perhaps there are some storage limits or such that I am not aware of, but this is still in a development phase and not rolled out yet and as I get the same result testing in the playground with the assistant I figured I’d report it.
I know this is similar to an existing topic, but that one gets a 500 error with ‘Failed to create file’, so I was a bit unsure if I should create a new topic or not. Let me know if I should move my post over there instead.