RAG Document QA System: Milvus, Cohere, OpenAI for Google Drive

This n8n workflow builds a Retrieval Augmented Generation (RAG) system that automatically processes new PDF files from a Google Drive folder. It triggers on new files, downloads them, extracts text content, chunks the data, generates embeddings using Cohere, and stores them in a Milvus vector database. This setup powers a conversational AI agent capable of answering queries by retrieving relevant document chunks and generating responses with OpenAI (ChatGPT 4o).

Key benefits include seamless automation of document ingestion for knowledge bases, cost-effective RAG with memory management for chat history, and scalability for large document sets. Ideal for teams needing instant QA over proprietary docs without manual indexing.

Use cases span enterprise search, customer support bots, legal research, or research assistants in health & fitness (e.g., querying nutrition PDFs). Adapt the chat trigger for webhooks, Slack, or apps. Requires API keys for Google Drive, Cohere, Milvus, and OpenAI.