Smarter RAG Agents: Enriched Retrieval & Modular AI Workflows

This n8n workflow provides an extendable RAG (Retrieval-Augmented Generation) template for creating powerful, explainable AI assistants. It handles file ingestion by uploading PDFs, extracting and chunking text, then embedding and storing in Supabase vector store using Google Gemini. Asynchronous enrichment adds LLM-generated metadata like topics, use cases, risks, audience level, and summaries to chunks, enabling filterable, metadata-based search. The agent chat pipeline processes user queries