Contextual RAG AI Agent with PGVector & Google Drive

This n8n workflow implements an advanced retrieval-augmented generation pattern using contextual embeddings for higher answer accuracy. It automatically pulls documents from a Google Drive folder, splits them into chunks, enriches each chunk with surrounding context before embedding, and stores everything in a PGVector-powered Postgres database. When users chat with the agent, it searches this contextually-aware knowledge base and combines it with conversation history to produce coherent, accurate responses. Ideal for teams building serious internal knowledge assistants who need better retrieval quality than basic chunk-based RAG setups provide.

Category: RAG & Knowledge Base