Hello,
I am trying to follow this community guide and I have a question: How to PDF into openAI (Solution!)
The assistant needs to be created based on the vector store or I can message a different assistant? I am wondering where I should input the specific guidelines and instructions for this agent and how to specify the information I want to extract the document
You set the assistant in open AI, use the playground to test this.
Then from make module you just call the assistant and pass it the information.
For example, you may say in the assistant, I want you to extract the top half of this pdf.
In the module, you can then just pass the pdf, no instructions, since the assistant has already been setup for this.
In short, set all your instructions in the openAI assistant.