{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/a90bed5b12434de3a098f3594e83e9bb\" frameborder=\"0\" width=\"1920\" height=\"1440\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1440,"width":1920,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1440,"thumbnail_width":1920,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/a90bed5b12434de3a098f3594e83e9bb-6036a80e9a55bb56.gif","duration":300.139,"title":"Building a RAG Q and A System","description":"This Loom explains a prototype for building a document Q&A RAG system for the reclaim process, using public APIs and Superbase to store and retrieve content. The author describes running a local backend and front end, ingesting documents into a vector store, and exposing the knowledge base via an MCP server tool, with citation-backed responses. They demonstrate a simple evaluation pipeline using FAFSA-related information and report position five out of twelve test question results with an overall precision of about 83% for 33%. They discuss next steps focused on improving reliability and measurable quality, including hybrid retrieval and re-ranking, benchmarking, and adding guardrails and better observability."}