home / mcp / supabase mcp server
Provides MCP-based CRUD access to Supabase data with safety checks, type validation, and flexible querying.
Configuration
View docs{
"mcpServers": {
"cloudmediaservices-supabase-mcp-server": {
"command": "python",
"args": [
"src/server.py"
],
"env": {
"SUPABASE_URL": "https://your-project-id.supabase.co",
"SUPABASE_SERVICE_ROLE_KEY": "your-service-role-key"
}
}
}
}You run a dedicated MCP server that lets AI assistants perform CRUD operations on your Supabase data with safety checks, type safety, and flexible querying. It supports reading, creating, updating, and deleting records across tables, with optional upsert, pagination, and filtering to empower intelligent workflows while protecting your data.
You connect an MCP client to the Supabase MCP Server and use its tools to perform data operations from your AI workflows. Start by running the local server, then configure your MCP client to point at the local process. You can issue read, create, update, and delete requests against any table in your Supabase project. Safety checks require confirmation for destructive actions, and you’ll receive detailed responses and errors to guide you.
Key capabilities you can leverage include reading table rows with filtering, creating new records, updating existing records based on conditions, and deleting records with safeguards. You can also combine operations with upsert behavior to insert or update in a single step, and tailor queries with pagination, ordering, and column selection.
Prerequisites: You need Python 3.11 or higher and a Supabase project with a service role key.
1) Install Python dependencies for the server.
2) Run the MCP server locally.
3) Configure your MCP client to connect to the server and supply your Supabase credentials.
Environment variables you must provide locally include the URL of your Supabase instance and the service role key. These credentials enable the server to access your database with proper permissions.
The server exposes a single, local process that you run with Python. Use environment variables to supply your connection details, and keep your keys secure.
Safety checks guard destructive operations. Input validation and comprehensive error handling help you diagnose issues quickly. Use the test suite to verify CRUD paths and edge cases.
Common issues include missing environment variables, invalid Supabase credentials, or permission problems with the service role key. Double-check your .env values and ensure the service role has access to the targeted tables.
The server provides a set of practical operations you can call from your MCP client to interact with Supabase data.
Query data from a table with support for filtering, ordering, and pagination.
Insert new records into a table with optional upsert behavior.
Update existing records using specified conditions.
Delete records with safety checks and explicit confirmation when required.