I am one of the devs who built this, and have a lot of opinions about RAG frameworks and constraints. Also knowledge graphs. So it would be great to see how anyone else has solved this or is working on the problem of specific retrieval for LLM output.
