information-retrieval-server MCP Server

A Model Context Protocol server for web information retrieval using Google Custom Search API and web crawling capabilities.

Features

Tools

• web_search - Perform web searches using Google Custom Search API
  • Returns titles, links, and snippets for top search results
  • Configurable result count (1-10)
• batch_web_search - Perform multiple web searches simultaneously
  • Execute up to 15 queries in parallel
  • Returns titles and links for each query
• google_image_search - Search for images using Google Custom Search
  • Returns image URLs, titles, and context
  • Configurable result count (1-10)
• web_crawl - Extract text content from webpages
  • Fast HTTP-based content extraction
  • Removes scripts, styles, and navigation elements
• advanced_web_crawl - Extract content using headless browser
  • More robust for JavaScript-heavy sites
  • Handles sites that block simple HTTP requests
  • Slower but more reliable

Development

Install dependencies:

npm install

Build the server:

npm run build

For development with auto-rebuild:

npm run watch

Installation

Quick Start (Recommended)

The easiest way to use this MCP server is with npx. No installation required!

On MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json On Windows: %APPDATA%/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "information-retrieval": {
      "command": "npx",
      "args": ["-y", "information-retrieval-mcp-server"]
    }
  }
}

Alternative: Global Installation

If you prefer to install the package globally:

npm install -g information-retrieval-mcp-server

Then configure:

{
  "mcpServers": {
    "information-retrieval": {
      "command": "information-retrieval-mcp-server"
    }
  }
}

Debugging

Since MCP servers communicate over stdio, debugging can be challenging. We recommend using the MCP Inspector, which is available as a package script:

npm run inspector

The Inspector will provide a URL to access debugging tools in your browser.