Hello 👋 . I am using a local llama LLM and doing RAG. Is there some way to retrieve the data in a structured format (JSON, dict, ...) without going through OpenAI?