In this video, Cole Medin presents an advanced Agentic RAG system that overcomes traditional RAG limitations by dynamically switching between vector-based document lookups and SQL-driven analysis of tabular data. The workflow includes modules for file extraction from Google Drive, intelligent data chunking, Postgres and Supabase integration, and embedding using OpenAI’s models. This comprehensive setup enables accurate source citation, improved context retrieval and detailed data analysis. The free template is available on GitHub.

Category: