Supabase Vector

Active
GitHub TypeScript Apache-2.0

Description

Supabase's built-in pgvector search, turning Postgres into a RAG database.

Key Features

  • pgvector — Native Postgres vector type and indexes
  • Hybrid search — Vector + full-text + scalar filters in a single SQL query
  • Realtime — Document changes pushed to clients via Realtime
  • RLS — Row-level security policies protect multi-tenant data
  • Edge Functions — Deno edge functions for embedding generation
  • Auth + Storage — Built-in auth and file storage

Use Cases

💡 Provide a one-stop backend plus vector search for SaaS apps.
💡 Quickly build a document RAG system with auth.
💡 Fuse traditional SQL business logic with vector search in the same DB.

Tags

vector-db postgres rag supabase

Categories

📚 RAG Tools

Quick Start

# Enable vector in SQL
create extension vector;
create table docs (id bigserial primary key, content text, embedding vector(1536));
# Query via the Supabase JS client
const { data } = await supabase.rpc('match_docs', {
  query_embedding: embedding, match_count: 5,
});
Visit GitHub

Related Projects