openai-cookbook/examples/vector_databases/supabase

Supabase Vector Database

Supabase is an open-source Firebase alternative built on top of Postgres, a production-grade SQL database.

Supabase Vector is a vector toolkit built on pgvector, a Postgres extension that allows you to store your embeddings inside the same database that holds the rest of your application data. When combined with pgvector's indexing algorithms, vector search remains fast at large scales.

Supabase adds an ecosystem of services and tools on top of Postgres that makes app development as quick as possible, including:

• Auto-generated REST APIs
• Auto-generated GraphQL APIs
• Realtime APIs
• Authentication
• File storage
• Edge functions

We can use these services alongside pgvector to store and query embeddings within Postgres.

OpenAI Cookbook Examples

Below are guides and resources that walk you through how to use OpenAI embedding models with Supabase Vector.

Guide	Description
Semantic search	Store, index, and query embeddings at scale using pgvector

Additional resources

• Vector columns
• Vector indexes
• RAG with permissions
• Going to production
• Deciding on compute