n8n Template: Notion → Supabase AI Knowledge Base

Turn your Notion knowledge into an AI-ready Supabase vector store—automatically

This n8n template syncs your entire Notion database into a Supabase vector store, chunking each page into embeddings so your content is ready for semantic search by an AI assistant or chatbot. Run it on a schedule to keep everything up to date—without manual cleanup or duplicate embeddings.

What this workflow does

• Clears the Supabase knowledge base before each run so the process is safe to repeat and always reflects the latest Notion content.
• Fetches all pages from your shared Notion database.
• Retrieves the full block content of each page.
• Skips empty pages to avoid wasting tokens and storage.
• Prepares documents for embedding by prepending each page title to its content and appending the source Notion URL.
• Splits content into chunks and generates embeddings using Ollama with nomic-embed-text.
• Stores embeddings in Supabase for semantic retrieval.

Use cases

• Build an AI knowledge base for customer support by searching across your Notion docs semantically.
• Enable internal copilots for SaaS teams to query product notes, policies, and runbooks stored in Notion.
• Keep documentation synced on a schedule so AI answers reflect the newest page content.

Technical details

• Integrations: Notion, Supabase (pgvector), Ollama
• Supabase setup: enable pgvector, create the documents table and match_documents function
• Workflow behavior: scheduled ingestion, safe re-runs via delete/clear step, chunking + embeddings, semantic search storage
• n8n nodes/components used: if, code, notion, sticky note, http request, manual trigger

Requirements: a Supabase account with pgvector enabled, a Notion internal integration shared to the target database, and Ollama running locally with nomic-embed-text.