What is the difference between trieve-type RAG and the new query tool? I am about to implement the new query tool and wondering what to expect.

Trieve type RAG is a search request against the files you upload where some chunks are returned and added to the context window. It's explaianable and around ~150ms in latency. The Query tool uses a LLM with your files in context to get a summary of what's relevant during a call. It's higher latency, ~750-1500ms but can be simpler to setup and a sane default.