RAG Knowledge Base Chatbot with OpenAI & MongoDB Vectors

This n8n workflow template consists of two interconnected automations for building a sophisticated knowledge base chatbot. Workflow 1 handles document ingestion and indexing: it pulls product documentation from Google Docs, splits it into searchable chunks, generates OpenAI embeddings, and stores them in MongoDB Atlas with metadata for semantic search. Workflow 2 powers the query interface: it embeds user questions, retrieves relevant chunks via vector similarity search in MongoDB, and uses GPT-4o-mini for retrieval-augmented generation (RAG) to produce precise responses, maintaining conversation context with a memory buffer.

Ideal for internal support teams, product specialists, and knowledge managers in tech companies, it solves the pain of manual document searches that cause delays and inconsistencies. By automating chunking, embedding, and retrieval, it enables instant, accurate AI-driven answers grounded in your proprietary docs, reducing support time and improving customer satisfaction.

Key benefits include scalability for large doc sets, cost-effective use of GPT-4o-mini, seamless integration with Google Workspace and MongoDB Atlas, and extensibility to webhooks or chat platforms. Use cases span product support Q&A, internal FAQ bots, training material access, and compliance querying, transforming static docs into dynamic knowledge assistants.