How can I use the API to upload a pdf as "knowledge" like in the web application?

I want to replicate the functionality of uploading a pdf, what is called “Knowledge” in the web application, using the API. I’m also curious if anyone might shed some light on using this feature in Azure’s GPT-4 API as well.

1 Like

Hey there and welcome to the community!

There’s some good info about this in the platform docs. You would be looking at “knowledge retrieval” under assistants, calling it as a tool.

https://platform.openai.com/docs/assistants/tools/knowledge-retrieval

2 Likes

When you look at the Assistants Playground you’ll notice it’s, essentially, a stripped down version of the CustomGPT interface.

TL;DR, this is exactly what you’re looking for.