BigQuery RAG with OpenAI for Documentation Q&A

This n8n workflow demonstrates a powerful RAG pipeline leveraging BigQuery's built-in vector capabilities and OpenAI models. It starts by embedding user queries with OpenAI, performs semantic search in a BigQuery table storing document chunks and embeddings (includes a public n8n docs table for instant testing), retrieves relevant context, and generates precise answers via OpenAI's LLM. No dedicated vector database needed—ideal for teams already using BigQuery.

Key benefits include cost-efficiency (leverages free BigQuery tier for small queries), seamless integration of enterprise data with OpenAI (bypassing direct compatibility issues), and scalability for production RAG apps. It reduces hallucinations by grounding responses in real docs, saving time on custom vector DB setups.

Use cases: Power AI chatbots for internal docs, customer support on product manuals, or knowledge bases in organizations with BigQuery data lakes. Perfect for n8n users exploring AI without extra infrastructure; extend to any document corpus by loading your own embeddings table.