Intelligent Document Q&A with AI and Vector Stores

This n8n workflow enables sophisticated question answering by integrating Large Language Models (LLMs) with a retrieval-augmented generation (RAG) approach. It listens for chat messages and uses an AI Agent to understand queries. For factual answers, it queries a Qdrant vector store, populated with embeddings generated by Ollama from your Google Drive documents. This ensures the AI has access to your specific knowledge base, providing accurate and relevant responses. The workflow also supports maintaining conversation history via a Postgres chat memory, allowing for more natural and contextually aware interactions. Documents added or updated in specified Google Drive folders are automatically processed to update the knowledge base.