josearubio/mcp-server-playground

MCP Servers Playground (Web Analyzer) 🛝

This is a sample MCP (Model Context Protocol) server built with Node.js and TypeScript.
It allows users to interact with websites through natural language queries using an MCP-compatible client (e.g., Claude).

The server supports detailed analysis of websites, including:

• Performance metrics
• SEO insights
• Content summaries and descriptions

This tool demonstrates how natural language interfaces can be integrated with web analysis services using the MCP protocol. In a way, what we are doing is providing context to LLMs, enabling them to query or perform operations that would not be possible by default in any chatbot or AI agent.

⚠️ Note: The current structure can definitely be improved — this was put together quickly to test integration with Claude Desktop.
Next steps include adding more tools and reorganizing the project to separate concerns properly (following a hexagonal architecture approach).

Features

• Integration with the Model Context Protocol (MCP).
• Analyze URLs for content, SEO, performance, and security (using Cheerio API). Content of web-scraped-service is AI generated, for testing purposes.
• HTTP API via Express (for API feature testing purposes only, not required for MCP configuration).

Tools

• analyze_url: Analyze URLs for content, SEO, performance, and security

Getting Started

Prerequisites

• Node.js >= 18
• npm
• Claude Desktop (or other MCP client)

Installation and run

npm install

npm run build

npm start

Usage

Add this configuration to your Claude Desktop configuration file:

(Usually at ~/.claude/claude_desktop_config.json on macOS/Linux or %APPDATA%\Claude\claude_desktop_config.json on Windows)

{
  "mcpServers": {
    "mcp-example": {
      "command": "node",
      "args": ["/path/to/your/project/dist/index.js"]
    }
  }
}

Start Claude Desktop and request to analyze a URL; it will ask for permission to use the tool.

To test via API:

curl -X POST http://localhost:3000/api/analyze \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com"}'