Store Notion Pages as Vectors in Supabase with OpenAI

This workflow triggers on new pages added to a Notion database, retrieves all block content, filters out media like images and videos, concatenates text into a single document, generates OpenAI embeddings, and stores everything with metadata in a Supabase vector table.

Benefits include seamless integration of Notion content into a vector database for RAG applications, LLMs, or semantic search without manual data handling. It saves hours of scripting and setup, ensuring scalable ingestion of documentation, notes, or articles.

Ideal use cases: Building AI-powered search over company wikis, automating knowledge base updates for chatbots, or creating personalized recommendation systems from Notion-stored content. Updated to summarize content and avoid duplicate rows.