🤖 MCP Integration

What is MCP?

The Model Context Protocol is an open standard allowing LLMs to interact with external tools. Beepack implements this protocol to enable AIs to:

• 🔍 Search for APIs by natural language description
• 📦 Install packages automatically
• 📖 Read package documentation
• ✅ Check compatibility with the current project

MCP Configuration

For Claude Desktop / OpenClaw

Add to your MCP configuration:

{
  "mcpServers": {
    "beepack": {
      "command": "beepack",
      "args": ["mcp-server"],
      "env": {
        "PACKBEE_TOKEN": "your-optional-token"
      }
    }
  }
}

For Cursor

In Cursor settings, add the Beepack MCP server:

// .cursor/mcp.json
{
  "servers": {
    "beepack": {
      "command": "beepack mcp-server"
    }
  }
}

Available Tools

search_packages

Search for packages by natural language description.

{
  "name": "search_packages",
  "description": "Search Beepack for reusable APIs matching a description",
  "parameters": {
    "type": "object",
    "properties": {
      "query": {
        "type": "string",
        "description": "Natural language description of what you need"
      },
      "capabilities": {
        "type": "array",
        "items": { "type": "string" },
        "description": "Specific capabilities required"
      },
      "compatible": {
        "type": "string",
        "description": "Runtime compatibility (cursor, copilot, claude)"
      }
    },
    "required": ["query"]
  }
}

Example AI usage:

// User asks: "Connect me to Notion"
// AI calls:
{
  "tool": "search_packages",
  "arguments": {
    "query": "sync with Notion database",
    "capabilities": ["read_database", "write_pages"]
  }
}

// Response:
{
  "packages": [
    {
      "slug": "notion-sync",
      "description": "Bidirectional sync with Notion",
      "stars": 45,
      "downloads": 1200,
      "score": 0.94
    }
  ]
}

install_package

Install a package in the current project.

{
  "name": "install_package",
  "description": "Install a Beepack package",
  "parameters": {
    "type": "object",
    "properties": {
      "slug": {
        "type": "string",
        "description": "Package slug (e.g., 'notion-sync')"
      },
      "version": {
        "type": "string",
        "description": "Specific version (optional, defaults to latest)"
      }
    },
    "required": ["slug"]
  }
}

get_package_info

Get package details.

{
  "name": "get_package_info",
  "description": "Get detailed information about a package",
  "parameters": {
    "type": "object",
    "properties": {
      "slug": {
        "type": "string",
        "description": "Package slug"
      }
    },
    "required": ["slug"]
  }
}

Example AI Workflow

// Scenario: User asks "Add Notion sync to my project"

// 1. AI searches first
const results = await mcp.call("search_packages", {
  query: "sync with Notion database"
});

// 2. Shows options to user
// "I found 'notion-sync' with 45 stars. Should I install it?"

// 3. On confirmation, installs
await mcp.call("install_package", {
  slug: "notion-sync"
});

// 4. Gets docs to guide user
const info = await mcp.call("get_package_info", {
  slug: "notion-sync"
});

// 5. Guides user
// "Package installed! You need to configure NOTION_API_KEY.
//  Here's how to use it: ..."

Benefits for AIs