undetected-chrome-mcp
Provides browser automation with enhanced anti-detection using undetected-chromedriver, enabling robust web scraping and interaction with sites that employ bot protection measures. It is fully compatible with server-puppeteer and includes network traffic monitoring capabilities.
README
Undetected ChromeDriver MCP Server
A Model Context Protocol (MCP) server providing browser automation capabilities using undetected-chromedriver for enhanced anti-detection compared to standard Puppeteer implementations.
๐ Features
- ๐ต๏ธ Enhanced Anti-Detection: Uses undetected-chromedriver to bypass modern anti-bot protections
- ๐ Drop-in Replacement: 100% API compatible with server-puppeteer
- โก High Performance: Optimized session management and resource cleanup
- ๐ก๏ธ Production Ready: Comprehensive error handling and logging
- ๐ฏ 14 Powerful Tools: Complete browser automation and network monitoring toolkit
- ๐ Network Traffic Monitoring: Capture and analyze all network requests/responses
๐ ๏ธ Available Tools
Core Browser Automation
| Tool | Description |
|---|---|
chrome_navigate |
Navigate to URLs with anti-detection |
chrome_screenshot |
Capture page or element screenshots |
chrome_click |
Click page elements |
chrome_fill |
Fill input fields |
chrome_select |
Select dropdown options |
chrome_hover |
Hover over elements |
chrome_evaluate |
Execute JavaScript code |
Network Traffic Monitoring
| Tool | Description |
|---|---|
start_network_monitoring |
Begin capturing network requests and responses |
stop_network_monitoring |
Stop network traffic capture |
get_network_requests |
Retrieve captured requests with filtering |
get_network_responses |
Get response data and headers |
get_response_body |
Extract specific response content |
get_network_summary |
Get network activity statistics |
clear_network_data |
Clear captured network data |
๐ฆ Installation
Prerequisites
- Python 3.8+
- Chrome/Chromium browser
- Git
Option 1: Install from PyPI (Recommended)
pip install undetected-chrome-mcp-server
Option 2: Install with uvx (Isolated Environment)
# Install with uvx for isolated execution
uvx --from undetected-chrome-mcp-server undetected-chrome-mcp-server
# Or run directly without installation
uvx undetected-chrome-mcp-server
Option 3: Install from Source
# Clone the repository
git clone https://github.com/andrewlwn77/undetected-chrome-mcp-server.git
cd undetected-chrome-mcp-server
# Create virtual environment (recommended for Docker)
python3 -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activate
# Install the package
pip install -e .
Dependencies
- mcp>=1.0.0 - Model Context Protocol
- undetected-chromedriver>=3.5.0 - Anti-detection Chrome automation
- selenium>=4.15.0 - WebDriver framework
- Pillow>=10.0.0 - Image processing for screenshots
โ๏ธ Configuration
Environment Variables
| Variable | Description | Default |
|---|---|---|
CHROME_EXECUTABLE_PATH |
Path to Chrome executable | /opt/google/chrome/google-chrome |
CHROME_DRIVER_LOG_LEVEL |
Logging level | INFO |
CHROME_SESSION_TIMEOUT |
Session timeout (seconds) | 300 |
CHROME_MAX_SESSIONS |
Max concurrent sessions | 5 |
CHROME_MAX_NETWORK_ENTRIES |
Max network requests/responses stored | 1000 |
MCP Configuration
Add to your .mcp.json:
Option 1: Using PyPI Installation
{
"mcpServers": {
"undetected-chrome": {
"command": "python",
"args": ["-m", "undetected_chrome_mcp.server"],
"env": {
"CHROME_EXECUTABLE_PATH": "/opt/google/chrome/google-chrome",
"CHROME_DRIVER_LOG_LEVEL": "INFO",
"CHROME_MAX_NETWORK_ENTRIES": "1000"
}
}
}
}
Option 2: Using uvx
{
"mcpServers": {
"undetected-chrome": {
"command": "uvx",
"args": ["undetected-chrome-mcp-server"],
"env": {
"CHROME_EXECUTABLE_PATH": "/opt/google/chrome/google-chrome",
"CHROME_DRIVER_LOG_LEVEL": "INFO",
"CHROME_MAX_NETWORK_ENTRIES": "1000"
}
}
}
}
Option 3: Using Virtual Environment (Docker/Development)
{
"mcpServers": {
"undetected-chrome": {
"command": "/path/to/project/venv/bin/python",
"args": ["/path/to/project/run_mcp_server.py"],
"cwd": "/path/to/project",
"env": {
"CHROME_EXECUTABLE_PATH": "/opt/google/chrome/google-chrome",
"CHROME_DRIVER_LOG_LEVEL": "INFO",
"CHROME_SESSION_TIMEOUT": "300",
"CHROME_MAX_SESSIONS": "5",
"CHROME_MAX_NETWORK_ENTRIES": "1000"
}
}
}
}
๐ Usage
Command Line
# Run the MCP server
python -m undetected_chrome_mcp.server
# Or use the installed command
undetected-chrome-mcp-server
Example Tool Calls
Navigate to a webpage:
{
"tool": "chrome_navigate",
"arguments": {
"url": "https://example.com",
"timeout": 30000
}
}
Take a screenshot:
{
"tool": "chrome_screenshot",
"arguments": {
"name": "example_page",
"encoded": true,
"fullPage": true
}
}
Execute JavaScript:
{
"tool": "chrome_evaluate",
"arguments": {
"script": "return document.title;"
}
}
๐ Network Traffic Monitoring
The undetected-chrome-mcp-server v1.1.0+ includes powerful network traffic monitoring capabilities using Chrome DevTools Protocol (CDP):
Basic Network Monitoring Workflow
// 1. Start monitoring
{
"tool": "start_network_monitoring"
}
// 2. Navigate and interact with pages
{
"tool": "chrome_navigate",
"arguments": {"url": "https://example.com"}
}
// 3. Get captured requests
{
"tool": "get_network_requests",
"arguments": {"url_filter": "api"}
}
// 4. Get response data
{
"tool": "get_network_responses",
"arguments": {"url_filter": "example.com"}
}
// 5. Extract specific response content
{
"tool": "get_response_body",
"arguments": {"request_id": "123.456"}
}
Key Features
- Full Traffic Capture: Captures all HTTP/HTTPS requests and responses
- Memory Protection: Configurable circular buffer (default: 1000 entries)
- Thread Safe: Concurrent access with proper locking
- Filtering: URL pattern filtering for requests/responses
- Complete Data: Headers, POST data, timestamps, status codes, response bodies
- Performance: Optimized for high-volume traffic capture
Use Cases
- API Data Extraction: Capture dynamic content loaded via AJAX/fetch
- Authentication Token Harvesting: Extract JWT tokens, session cookies
- Performance Analysis: Monitor request timing and response sizes
- Debugging: Inspect network calls for troubleshooting
- Data Mining: Collect structured data from API endpoints
๐งช Testing
The project includes test scripts to validate functionality:
# Available test scripts in root directory:
python test_functions.py # Basic function validation
python test_reddit_automation.py # Reddit anti-detection tests
python test_undetected_chrome.py # Core driver tests
Test Results
- โ Navigation: Successfully loads protected pages
- โ JavaScript Execution: Full DOM access and data extraction
- โ Screenshots: High-quality image capture
- โ Anti-Detection: Bypasses modern bot protection (tested on Reddit)
๐๏ธ Architecture
The server follows a layered architecture:
- MCP Protocol Layer - Handles MCP communication
- Browser Management Layer - Manages Chrome sessions
- Anti-Detection Layer - Implements stealth capabilities
- Operation Execution Layer - Executes browser operations
๐ก๏ธ Anti-Detection Features
- Undetected ChromeDriver: Advanced stealth automation
- User Agent Rotation: Randomized browser fingerprints
- Viewport Randomization: Human-like window sizing
- Human-like Delays: Natural interaction timing
- JavaScript Stealth: Removes automation indicators
- Session Isolation: Clean state between operations
๐ฏ Comparison with server-puppeteer
| Feature | server-puppeteer | undetected-chrome-mcp |
|---|---|---|
| Anti-Detection | Basic | Advanced |
| Bot Protection Bypass | Limited | Excellent |
| API Compatibility | N/A | 100% Compatible |
| Performance | Good | Optimized |
| Resource Management | Basic | Advanced |
๐ค Contributing
- Fork the repository
- Create a feature branch:
git checkout -b feature-name - Commit changes:
git commit -am 'Add feature' - Push to branch:
git push origin feature-name - Create a Pull Request
๐ License
This project is licensed under the MIT License - see the LICENSE file for details.
๐ Acknowledgments
This project builds upon excellent open source work:
- server-puppeteer - Original MCP Puppeteer server that inspired this implementation
- undetected-chromedriver - The core anti-detection technology that powers this server
- Model Context Protocol (MCP) - The foundation protocol enabling AI-tool integration
Special thanks to the open source community for creating these foundational tools that make advanced browser automation accessible to everyone.
๐ Support
- GitHub Issues: Report bugs or request features
- Examples: Check the test scripts in root directory for usage examples
๐ Related Projects
- reddit-mcp - Reddit research MCP server
- Model Context Protocol - Official MCP documentation
Built with โค๏ธ by andrewlwn77
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.