Build Documentation Expert Chatbot with Gemini RAG Pipeline

This n8n workflow implements a complete Retrieval-Augmented Generation (RAG) pipeline using Gemini AI. It consists of two main parts: a one-time indexing process that scrapes documentation (e.g., n8n docs), chunks content, generates embeddings, and stores them in n8n's Simple Vector Store; and a chat interface where user queries retrieve relevant chunks to feed into Gemini, ensuring factual responses based solely on provided knowledge.

Benefits include creating specialized AI experts without hallucinations, ideal for internal knowledge bases. It's efficient for onboarding, support, or training, saving hours of manual research. The in-memory store is temporary, requiring re-indexing after restarts, but setup is quick (~2 min + 15-20 min indexing).

Use cases: HR onboarding chatbots for employee handbooks, operations manuals Q&A, customer support on product docs, or any domain-specific expertise. Adaptable to any URL-scrapable docs, leveraging n8n's built-in tools for scalability.

High business value in AI-driven automation, reducing reliance on general LLMs and enabling precise, auditable answers.