Browser Tools MCP Server

Powerful browser automation, auditing, and debugging tools for AI assistants - control browsers, capture screenshots, analyze performance, and audit web applications

View on GitHub Quick Start →

⚡ Quick Install with Claude Code MCP Guide

Connect Claude Code to any remote MCP server with one simple command. Choose between direct HTTP connections or a universal bridge - all configured automatically with our setup wizard.

Get the Claude Code MCP Setup Guide →

🔌 Test MCP Server Connection

Test your connection to the MCP wrapper server using your x-api-key. This validates that your MCP server credentials are configured correctly.

MCP Server API Key (x-api-key)

🚀 Test Full Browser Connection

Test both your MCP server connection AND your browser configuration. This ensures your complete browser automation setup is working correctly.

MCP Server API Key (x-api-key)

Browser Configuration (x-browser-config)

📡 API Endpoint & Required Headers

POST https://browser-tools.mcp.majewscy.tech

Required HTTP Headers:

Content-Type: application/json

x-api-key: Your MCP server API key

x-browser-config: Browser configuration JSON (optional)

🛠️ Available Tools

take_screenshot

Capture screenshots of web pages or specific elements

navigate_to

Navigate browser to any URL

get_console_logs

Retrieve browser console logs and errors

run_lighthouse_audit

Perform Google Lighthouse performance audits

click_element

Click on page elements using selectors

fill_form

Fill out web forms automatically

extract_text

Extract text content from page elements

analyze_network

Monitor and analyze network requests

⚠️ Common Errors & Solutions

401 Unauthorized

Invalid API credentials

Solution: Check your x-api-key is correct

Browser Launch Failed

Unable to start browser instance

Solution: Ensure browser dependencies are installed and server has sufficient resources

Navigation Timeout

Page failed to load within timeout period

Solution: Check URL is accessible and increase timeout in browser configuration

Element Not Found

Specified CSS selector did not match any elements

Solution: Verify the CSS selector is correct and element is visible on the page

🌐 Browser Configuration

This MCP server uses Puppeteer to control a headless Chrome browser. The server handles browser management automatically, but you can customize browser behavior through configuration.

💡 Note: The server runs a sandboxed browser instance with security restrictions for safety. Some features may be limited in headless mode.

Browser Configuration Example

{
  "headless": true,
  "defaultViewport": {
    "width": 1920,
    "height": 1080
  },
  "args": [
    "--no-sandbox",
    "--disable-setuid-sandbox",
    "--disable-dev-shm-usage"
  ],
  "timeout": 30000,
  "waitForSelector": 5000,
  "screenshotOptions": {
    "type": "png",
    "fullPage": false,
    "quality": 90
  }
}

Configuration Options:

headless: Run browser in headless mode (true/false)
defaultViewport: Set browser window dimensions
timeout: Page load timeout in milliseconds
waitForSelector: Element wait timeout in milliseconds
screenshotOptions: Screenshot format and quality settings

🏥 Service Health Check

Checking...

Response Time

Server Status

Protocol Version

2024-11-05

Last Checked

Configuration

// Claude Desktop: ~/Library/Application Support/Claude/claude_desktop_config.json (macOS)
// Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "browser-tools": {
      "command": "npx",
      "args": [
        "-y",
        "@agentdeskai/browser-tools-mcp@latest"
      ]
    }
  }
}

// HTTP Wrapper Configuration (for remote access)
{
  "browser-tools": {
    "type": "http",
    "url": "https://browser-tools.mcp.majewscy.tech",
    "headers": {
      "x-api-key": "YOUR_MCP_API_KEY"
    }
  }
}

// Claude Code: .mcp.json in your project root (for HTTP wrapper)
{
  "browser-tools": {
    "type": "http",
    "url": "https://browser-tools.mcp.majewscy.tech",
    "headers": {
      "x-api-key": "YOUR_MCP_API_KEY"
    }
  }
}

Features

Comprehensive browser automation and web analysis tools for AI assistants

📸

Screenshots

Capture full page or element-specific screenshots with customizable options.

🎯

Element Interaction

Click, fill forms, and interact with web elements using CSS selectors.

📊

Lighthouse Audits

Run performance, accessibility, and SEO audits using Google Lighthouse.

🔍

Console Logs

Monitor JavaScript errors, warnings, and console output.

🌐

Network Analysis

Track HTTP requests, responses, and network performance.

📱

Device Emulation

Test websites on different devices and screen sizes.

📋 Recent Updates from GitHub

Latest Aug 29, 2025

Feature: Enhanced screenshot capabilities with element targeting
Fix: Improved browser instance lifecycle management

Recent Aug 28, 2025

Feature: Added Lighthouse audit integration
Enhancement: Better error handling for network timeouts

Recent Aug 27, 2025

Feature: Console log monitoring and filtering
Fix: Resolved element selector timeout issues

Recent Aug 26, 2025

Enhancement: Improved form filling capabilities
Feature: Network request/response monitoring

Recent Aug 25, 2025

Feature: Added text extraction from page elements
Fix: Browser configuration parsing improvements

Recent Aug 24, 2025

Feature: Initial browser automation framework
Docs: Added comprehensive API documentation

View all commits on GitHub →

🔐 Update Server