Build & Query RAG System with Google Drive, OpenAI GPT-4o-mini, Pinecone

This n8n workflow creates a complete Retrieval-Augmented Generation (RAG) system that automatically ingests documents from a specified Google Drive folder. Upon detecting new files, it downloads them, splits text into chunks using a recursive splitter, loads them as LangChain documents, generates embeddings with OpenAI, and upserts them into a Pinecone vector database under the 'ragfile' index. This ingestion path runs every minute, ensuring real-time updates to your knowledge base.

The query path powers a chat interface where user messages are processed by a LangChain AI Agent using GPT-4o-mini. The agent retrieves relevant document chunks from Pinecone based on query embeddings, providing context-aware responses. This setup leverages advanced AI for accurate, grounded answers without hallucinations.

Benefits include seamless automation of knowledge base management, scalability for large document sets, and cost-effective use of GPT-4o-mini. Ideal use cases: customer support chatbots with proprietary docs, internal company wikis, research assistants querying reports, or legal/contract review systems. Requires connecting Google Drive, OpenAI, and Pinecone credentials in n8n.