🤖 Documentation Expert Bot with RAG, Gemini & Supabase

This workflow creates a complete Retrieval-Augmented Generation (RAG) pipeline to transform the n8n documentation into an expert AI chatbot. It scrapes official n8n docs, chunks content, generates embeddings, and stores them in a Supabase vector database for efficient retrieval. The chat interface then queries this knowledge base to fetch relevant chunks and feeds them to Gemini, ensuring responses are factual and grounded in the source material.

Part 1 focuses on one-time indexing: automatically scrape all n8n documentation pages, break them into manageable chunks, embed them using AI, and index in Supabase. This builds a private, searchable knowledge base like a hyper-detailed librarian's catalog.

Part 2 powers the interactive chatbot: user questions trigger similarity search in the vector store, retrieve top relevant chunks, and prompt Gemini to answer solely based on that data, minimizing hallucinations.

Benefits include hyper-accurate AI responses for technical docs, easy customization to any documentation set, and significant time savings for support teams or developers. Use cases: internal knowledge bots, customer support automation, training assistants, or expert systems for any domain-specific content.