Using OpenAI Responses API: Web Search, RAG & More

So if we pass the response id and don't pass the previous convo then it won't consume extra tokens right?

I love this concept. I need to upload a pdf document and parse it with PyMuPDF then pull out the data with a defined structure. Any suggestion where I need to look for help with the code. I am a big user of Claude to write code but the Responses structure is so new that no AI model can tweak the code.

can we get GitHub link for this features

What about the cost of the websearch? Looks like $0.025 per query. Completely unviable for a production application with lots of users. I created an app and this seemed to be reflected in the costs. 1000 users * 10 queries a day * 0.025 * 30 days = $7500 / month. Am I reading this right?

Great presentation! Thank you Ben!

Is code available somewhere?