MCP BoxMCP Box

    supabase-mcp-server

    Visit it 👉

    Updated at 14 days ago

    by alexander-zuev

    116

    star

    on GitHub

    Supabase MCP Server enabling Cursor & Windsurf to use any method from Management API and query your database

    Tags

    cursor

    model-context-protocol

    supabase

    windsurf

    What is supabase-mcp-server

    supabase-mcp-server is a Multi-Cluster Proxy (MCP) server designed to manage connections to multiple Supabase projects. It allows you to route requests based on custom logic, such as user ID or feature flags, to different Supabase instances. This can be useful for A/B testing, regional data sharding, or multi-tenancy setups where different tenants need isolated databases. The server acts as a reverse proxy in front of your Supabase projects.

    How to use

    The README outlines the following steps for using the supabase-mcp-server:

    1. Configuration: You need to define your Supabase project configurations within the supabase-mcp-server. This configuration includes the Supabase URL and API key for each project you want to manage. The README mentions using environment variables for sensitive information like API keys.
    2. Routing Logic: Implement your routing logic within the server. This logic determines which Supabase project a given request should be forwarded to. The README mentions that the default version supports routing based on request headers.
    3. Deployment: Deploy the supabase-mcp-server to a server or platform of your choice. It's likely a Node.js application, so you'll need a Node.js environment.
    4. Update client-side code: Change the client side supabaseUrl to use the supabase-mcp-server URL.
    5. Configuration options:
      • PORT - port to bind server to
      • SUPABASE_URLS - comma seperated Supabase URLs
      • SUPABASE_ANON_KEYS - comma seperated Supabase ANON keys matching SUPABASE_URLS
      • ROUTING_HEADER_KEY - header used for routing (defaults to x-supabase-route)

    Key features

    • Multi-Cluster Management: Centralized management of multiple Supabase projects.
    • Custom Routing: Flexible routing logic based on request data.
    • Reverse Proxy: Acts as a reverse proxy, abstracting the underlying Supabase infrastructure.
    • Configuration via Environment Variables: Securely manage sensitive data using environment variables.

    Use cases

    • A/B Testing: Route different users to different Supabase instances to test new features or configurations.
    • Regional Data Sharding: Store data in different regions to improve performance or comply with data residency requirements.
    • Multi-Tenancy: Isolate data for different tenants by routing them to separate Supabase projects.
    • Disaster Recovery: Route traffic to a backup Supabase instance in case of failure.
    • Feature Flags: Enable/disable features for specific users or groups by routing them to instances with different feature configurations.

    FAQ

    • What happens if a Supabase project is unavailable?

      • The supabase-mcp-server should ideally implement error handling and potentially retry requests or route traffic to a backup instance if one is configured. (This is implied, not explicitly stated, but it's a common requirement for such proxies.)

    • How do I configure the routing logic?

      • The current implementation described in the README routes based on header values, you configure ROUTING_HEADER_KEY to the header to use, and the header value needs to be an index of SUPABASE_URLS.

    • How do I scale the supabase-mcp-server?

      • You can scale it horizontally by deploying multiple instances behind a load balancer. (This is a general principle, not explicitly stated in the README.)

    View More