If you want to learn:
How do I set up Supabase database tables for n8n RAG integration?
What is the correct database schema for storing vector embeddings in Supabase?
How do I configure PostgreSQL extensions for vector storage in Supabase?
What are embedding dimensions and why do they matter for RAG systems?
How do I create custom SQL functions for n8n's vector store operations?
What's the proper way to structure a knowledge base table for retrieval-augmented generation?
Then this lecture is for you!
In this hands-on lecture, you'll learn how to configure Supabase as a vector store for your n8n RAG workflow. You'll start by navigating the Supabase dashboard and understanding the project structure. The lecture walks you through enabling the vector extension in your PostgreSQL database, which is essential for storing embeddings and performing vector search operations.
You'll execute custom SQL code to create a knowledge_base table with the proper schema that n8n expects, including fields for id, content, metadata, and embedding vectors. The lecture explains how to set up the critical match_documents function that n8n uses to retrieve relevant context from your vector database during the RAG process.
A key focus is understanding embedding dimensions and why they must match your chosen embedding model. You'll learn why OpenAI's Embedding Small model uses 1,536 dimensions and how to configure your Supabase database accordingly. The lecture also covers alternative embedding options including Gemini and open-source models.
By the end, you'll have a fully configured Supabase vector database ready to store documents from Google Drive and support your AI agent's retrieval operations. You'll verify the successful table creation through the Supabase interface and prepare for the next step: connecting n8n to populate your knowledge base with automated file ingestion from Google Drive.