Introduction to MCP

The Model Context Protocol (MCP) is a standardized way for AI assistants (like Cursor, Claude, Windsurf) to connect to external data sources and tools.

Think of it as adding "plugins" to your AI. Instead of just knowing about your code files, your AI can now directly access your:

• Database (Supabase, Firebase)
• Design Files (Figma)
• Subscription Data (RevenueCat)
• Documentation (Context7)

Why is this important?

Without MCP, your AI is blind to your data.

• ❌ It doesn't know your database schema.
• ❌ It can't see your design specs.
• ❌ It guesses column names and often gets them wrong.

With MCP, your AI becomes context-aware:

• ✅ Schema Awareness: It reads your actual database structure.
• ✅ Real-time Data: It can query recent users or logs to debug issues.
• ✅ Design-to-Code: It can inspect Figma nodes to generate pixel-perfect UI.

How it Works

You run a lightweight "MCP Server" locally (or via npx). Your AI client (Cursor/Claude) connects to this server.

graph LR
    A[AI Assistant] -->|MCP Protocol| B[MCP Server]
    B -->|Query/Action| C[External Service]
    C -->|Data| B
    B -->|Context| A

Supported Servers

We have pre-configured support for the following MCP servers. Click to learn how to set them up:

Best Practices

Security and performance are critical when giving AI access to your infrastructure.