Supabase MCP

Official Supabase MCP server — manage your Supabase projects, run SQL, handle migrations, and interact with Edge Functions via AI.

TRUSTED · Official by Supabase

Database APIs

// Add to your client

{
  "mcpServers": {
    "supabase": {
      "command": "npx",
      "args": ["-y", "@supabase/mcp-server-supabase@latest", "--access-token", "<your-personal-access-token>"]
    }
  }
}

{
  "mcpServers": {
    "supabase": {
      "command": "npx",
      "args": ["-y", "@supabase/mcp-server-supabase@latest", "--access-token", "<your-personal-access-token>"]
    }
  }
}

claude mcp add supabase -- npx -y @supabase/mcp-server-supabase@latest --access-token <your-personal-access-token>

Paste into your client's MCP configuration file.

// Try it now

"List all tables in my Supabase database and show their row counts"

! Use a project access token scoped to the right project — not your personal account token
! Running migrations on production is destructive — always review generated SQL before applying
! RLS policies can hide data from the MCP server — query results may look incomplete

Setup

Easy Setup

API key or simple config

Pricing

Free

Free for Supabase free tier. Your Supabase plan costs apply separately.

Token usage

Moderate usage

// When to use

You're building an app on Supabase and want AI help with schema and queries
You need to inspect production or staging data without leaving your editor
You want AI to draft and apply migrations based on conversation
You're debugging auth issues and want to query the auth schema

// When NOT to use

You use Firebase, PlanetScale, or self-hosted Postgres (use those specific servers)
You want to give AI write access to production without a review gate
You need real-time WebSocket subscriptions from the MCP server (not supported)

// Usage Scenarios

Schema Design with AI

Discuss a feature and let AI generate and apply the migration.

Example prompt

"Add a 'tags' jsonb column to the posts table with a GIN index and create the migration"

Data Investigation

Run ad-hoc queries to debug issues or understand user behavior.

Example prompt

"Show me the 10 users who signed up most recently and their completed onboarding steps"

Edge Function Deployment

Write, deploy, and test Edge Functions through conversation.

Example prompt

"Create an Edge Function that sends a welcome email when a new user signs up"

// About

Plain English

Imagine telling your AI: 'Show me the last 100 signups from the users table' or 'Create a new migration that adds a created_at column' — and it manages your Supabase backend directly. This server connects your AI to Supabase so it can query databases, run migrations, deploy Edge Functions, and manage auth and storage.

The official Supabase MCP server gives AI assistants comprehensive access to the Supabase platform. Manage databases, run SQL queries, handle migrations, configure authentication, manage storage buckets, and interact with Edge Functions — all through natural language.

Maintained by the Supabase team with deep integration into the Supabase CLI and dashboard.

GitHub 3.2kUpdated Feb 14, 2025Apache-2.0

View on GitHub Official Site