I am trying to have an RAG chat using Pinecone to vector the data, but it doesn't seem like I am getting the expected results. I have about 1.2 million docs to vector, and I currently have about 4k, but the level of connection is not very good. I would love to chat with anyone who knows the flow. I really appreciate any help you can provide.

I was struggling with generating diagrams (technical diagrams) so i decided to create a GPT which can do that for me i am pretty happy with this. Any thoughts? It can also pull transcripts from youtube and visualise them with summary and time stamps. https://chat.openai.com/g/g-7EWovgPuJ-mindstream