Back to MCP Catalog

Web Research MCP Server

Browser AutomationTypeScript
Enables Claude to perform real-time web research with Google search and content extraction
Available Tools

search_google

Performs Google searches and extracts results

query

visit_page

Visits a webpage and extracts its content

urltakeScreenshot

take_screenshot

Takes a screenshot of the current page

Web Research is a powerful tool that brings real-time internet access to Claude, allowing it to search the web, visit pages, and extract relevant information. It maintains a research session that tracks visited pages, search queries, and captured screenshots, making it ideal for comprehensive research tasks. With Web Research, Claude can perform Google searches, navigate to websites, extract content, and even take screenshots of pages. This enables more informed responses based on current information rather than relying solely on Claude's training data. The integration is designed to support thorough research workflows with features for tracking sources and maintaining context throughout a research session.

Overview

Web Research is a Model Context Protocol (MCP) server that enables Claude to perform real-time web research. It allows Claude to search the web, visit webpages, extract content, and capture screenshots, making it an invaluable tool for research tasks that require current information.

Prerequisites

Before installing Web Research, ensure you have:

Installation

To install Web Research, you need to add it to your Claude Desktop configuration:

  1. Locate your claude_desktop_config.json file:

    • On macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
    • On Windows: %APPDATA%\Claude\claude_desktop_config.json
  2. Add the following entry to the configuration file:

{
  "mcpServers": {
    "webresearch": {
      "command": "npx",
      "args": ["-y", "@mzxrai/mcp-webresearch@latest"]
    }
  }
}
  1. Save the file and restart Claude Desktop if it's already running.

This configuration allows Claude Desktop to automatically start the Web Research MCP server when needed.

Using Web Research

You can use Web Research in two ways:

Method 1: Using the Agentic Research Prompt

  1. Start a new chat in Claude Desktop
  2. Click the Paperclip icon in the chat input
  3. Select "Choose an integration" → "webresearch" → "agentic-research"
  4. The prompt will guide Claude to conduct thorough web research on your topic

Method 2: Direct Interaction

Simply start a chat with Claude and ask questions that would benefit from web research. Claude will use the available tools to search for information and visit relevant webpages.

Research Session

Web Research maintains a research session that includes:

  • Search queries performed
  • Webpages visited
  • Content extracted from pages
  • Screenshots captured
  • Timestamps of activities

You can access captured screenshots in Claude Desktop via the Paperclip icon.

Tips for Better Results

  • When researching general topics, suggest high-quality sources for Claude to use (e.g., "news today from Reuters or AP" instead of just "news today")
  • Be specific in your research questions to help Claude focus the search
  • Ask Claude to cite sources with URLs when providing information
  • For complex research tasks, break them down into smaller, more focused questions

Troubleshooting

If you encounter issues with Web Research:

  1. Check Claude Desktop's MCP logs:

    tail -n 20 -f ~/Library/Logs/Claude/mcp*.log
    
  2. Ensure you have the correct Node.js version installed (18+)

  3. Try restarting Claude Desktop

  4. If problems persist, you may need to manually install the package:

    npm install -g @mzxrai/mcp-webresearch
    

Limitations

  • Web Research is pre-alpha software and may contain bugs
  • Some websites with strong bot detection might block access
  • Complex web applications or sites requiring authentication may not work properly
  • Currently verified primarily on macOS, with limited testing on other platforms

Related MCPs

Playwright Browser Automation
Browser AutomationPython

Automate browser interactions with Playwright

Playwright Browser Automation
Browser AutomationJavaScript

Automate browser interactions, take screenshots, and scrape web content

Playwright Browser Automation
Browser AutomationTypeScript

Browser automation capabilities using Playwright

About Model Context Protocol

Model Context Protocol (MCP) allows AI models to access external tools and services, extending their capabilities beyond their training data.

Generate Cursor Documentation

Save time on coding by generating custom documentation and prompts for Cursor IDE.