Automated Knowledge Base Query with OpenAI and Notion

This workflow facilitates dynamic interaction with a knowledge base stored across Notion and a vector database, leveraging OpenAI’s language models. It enables scheduled or real-time question answering by retrieving and embedding page content, storing it in a vector database, and then querying it with user prompts. Key nodes include Notion for fetching page blocks, OpenAI for embeddings and chat responses, and Supabase as a vector store. The workflow deletes outdated embeddings, processes page content into manageable chunks, and maintains a conversational context for user inquiries. This setup is especially useful for creating intelligent chatbots that can answer questions from a constantly updated knowledge repository, making it ideal for support, documentation, and knowledge management scenarios.