What it does
This n8n workflow watches a Notion database for new pages and reads their text blocks, leaving out media blocks such as images and videos. It joins the remaining text, creates embeddings with OpenAI and stores each document with its vector in a Supabase table that has a vector column. That table must exist before you start.
Use cases
- 01Index meeting notes added to a Notion database
- 02Make team wiki pages searchable by meaning
- 03Feed the retrieval step of an internal assistant
Connects
Notion · Embeddings OpenAI · Token Splitter · Supabase Vector Store · Default Data Loader