Supabase
by Supabase Community
The Supabase MCP Server connects your Supabase projects to AI assistants like Claude, Cursor, and Windsurf through the Model Context Protocol standard. It exposes 34 tools across eight feature groups — account management, documentation search, database operations, debugging, development utilities, Edge Functions, branching, and storage — giving LLMs direct access to query databases, apply migrations, deploy Edge Functions, manage branches, and monitor logs. Built by the official Supabase community organization, this server supports both cloud-hosted and local deployment modes. The cloud version at mcp.supabase.com uses OAuth 2.1 for seamless authentication, while local installations work through personal access tokens. Configuration is granular: developers can enable read-only mode to prevent write operations, scope the server to a single project to limit blast radius, and selectively enable or disable feature groups to minimize the exposed tool surface. Security is a first-class concern. SQL query results are wrapped with prompt injection mitigations, and the server recommends connecting to development projects rather than production databases. The branching system allows safe experimentation by creating isolated database copies before merging changes to production. Currently in public alpha with the npm package @supabase/mcp-server-supabase, the server sees active development with approximately 50,000 total npm downloads and growing adoption across the AI-assisted development ecosystem.
Installation
Key Features
- ✓34 MCP tools across 8 feature groups: account, docs, database, debugging, development, Edge Functions, branching, and storage
- ✓Execute raw SQL queries and apply database migrations directly through AI assistants
- ✓Deploy and manage Supabase Edge Functions without leaving your AI coding environment
- ✓Read-only mode prevents all write operations for safe data exploration
- ✓Project scoping restricts access to a single Supabase project to limit blast radius
- ✓Built-in prompt injection mitigations wrap SQL results with safety instructions
Use Cases
- →Scaffold a full Next.js application with Supabase backend using AI-assisted development in Cursor or Claude Code
- →Debug production issues by querying logs and running diagnostic SQL through your AI assistant
- →Generate TypeScript types from your database schema automatically during development
- →Create and test database migrations on isolated branches before merging to production
- →Deploy Edge Functions directly from AI conversations without switching to the Supabase dashboard