Home - JinsongRoh/pydoll-mcp GitHub Wiki

πŸ€– PyDoll MCP Server Wiki

PyDoll Logo

The Ultimate Browser Automation MCP Server
Revolutionary Zero-WebDriver Automation with Intelligent CAPTCHA Bypass

πŸ“– Wiki Table of Contents

πŸš€ Getting Started

πŸ“š User Guide

πŸ”§ Advanced Features

πŸ› Troubleshooting

πŸ”„ Updates and Development


πŸ“‹ PyDoll MCP Server - Project Introduction and Overview

πŸš€ Project Basic Information

Overview

PyDoll MCP Server (pydoll-mcp) v1.1.3 is a Model Context Protocol (MCP) server built on the PyDoll library. It enables AI assistants like Claude, OpenAI, and Gemini to naturally control web browsers and perform complex web automation tasks.

Core Features & Revolutionary Technology

  • 🚫 Zero WebDriver: Direct browser communication through Chrome DevTools Protocol
  • 🧠 AI-powered CAPTCHA Bypass: Automatic resolution of Cloudflare Turnstile and reCAPTCHA v3
  • πŸ‘€ Human Behavior Simulation: Undetectable interactions that bypass sophisticated anti-bot systems
  • ⚑ Native Async Architecture: Ultra-fast concurrent automation
  • πŸ•΅οΈ Advanced Stealth Mode: Anti-detection technology that makes automation undetectable

User-Friendly

  • πŸ”§ One-Click Setup: Automatic Claude Desktop configuration
  • 🌍 Universal Compatibility: Works on all systems including Korean Windows
  • πŸ› MCP Protocol Compliance: Resolved JSON parsing issues for stable communication
  • πŸ“š Rich Documentation: Complete user guide and API reference

πŸ”§ Technical Details

Foundation Technology

  • Based on Python 3.8+
  • Utilizes PyDoll 2.2.1 (2025.06.17) library
  • Supports Model Context Protocol (MCP)
  • Browser control based on Chrome DevTools Protocol

Major Feature Categories

  • Browser start/stop, tab management, status monitoring, etc.

🧭 Navigation and Page Control (10 tools)

  • URL navigation, page refresh, back/forward navigation, page load waiting, etc.

🎯 Element Finding and Interaction (15 tools)

  • Element search, clicking, text input, keyboard input, drag and drop, etc.

πŸ“Έ Screenshots and Media (6 tools)

  • Page/element screenshots, PDF generation, video capture, etc.

⚑ JavaScript and Advanced Scripting (8 tools)

  • JavaScript execution, script injection, console log monitoring, etc.

πŸ›‘οΈ Protection Bypass and Stealth (12 tools)

  • Cloudflare/reCAPTCHA bypass, stealth mode, bot challenge handling, etc.

🌐 Network Control and Monitoring (10 tools)

  • Network monitoring, request interception, API response extraction, etc.

πŸ“ File and Data Management (8 tools)

  • File upload/download, data extraction/export, etc.

🎯 Main Use Cases

Smart Web Automation

  • Navigate websites with human-like behavior patterns
  • Extract data from protected dynamic websites
  • Automate complex workflows across multiple pages
  • Seamlessly handle modern SPAs and dynamic content

Protection System Bypass

  • Automatically solve Cloudflare Turnstile CAPTCHAs
  • Bypass reCAPTCHA v3 without external services
  • Evade sophisticated bot detection systems
  • Navigate protected content areas

Advanced Data Extraction

  • Scrape data from modern protected websites
  • Monitor and capture all network API calls
  • Extract information from dynamic, JavaScript-heavy sites
  • Handle complex authentication flows

πŸ“Š Performance Metrics

Item PyDoll MCP Traditional Tools
Setup Time < 30 seconds 5-15 minutes
CAPTCHA Success Rate 95%+ 20-30%
Detection Evasion 98%+ 60-70%
Memory Usage 50% less Baseline
Speed 3x faster Baseline
Reliability 99%+ 80-85%

πŸš€ Quick Start, πŸ”§ Setup and Integration

Installation Method

pip install pydoll-mcp

Claude Desktop Auto Setup

From v1.1.0, automatic Claude Desktop configuration is provided after installation:

python -m pydoll_mcp.cli auto-setup

Manual Setup

Add the following to your Claude Desktop configuration file:

{
  "mcpServers": {
    "pydoll": {
      "command": "python",
      "args": ["-m", "pydoll_mcp.server"],
      "env": {
        "PYDOLL_LOG_LEVEL": "INFO"
      }
    }
  }
}

πŸ†• Latest Updates (v1.1.3 - 2025-06-18)

πŸ› Critical Bug Fixes

  • βœ… JSON Parsing Error Fixed: Resolved MCP client communication issues
  • βœ… Encoding Compatibility: Full support for Korean Windows systems (CP949/EUC-KR)
  • βœ… Protocol Compliance: Proper stdout/stderr separation for MCP compatibility
  • βœ… Enhanced Stability: Improved server startup and error handling

πŸ“š Related Resources

πŸ“ Latest Releases

v1.1.3 (2025-06-18)

  • Resolved JSON parsing errors
  • Korean Windows system support
  • Improved MCP protocol compliance
  • Enhanced stability and error handling

v1.1.2

  • Enhanced server stability
  • Performance optimization
  • Reduced memory usage

v1.1.1

  • Fixed Korean Windows environment issues
  • Improved international compatibility
  • Added auto-recovery features

v1.1.0

  • One-click auto setup feature
  • Enhanced CLI tools
  • Improved developer experience

🀝 Community

πŸ“„ License

This project is distributed under the MIT License. For details, see the LICENSE file.


Ready to revolutionize browser automation?
Installation Guide | Quick Start | API Reference

PyDoll MCP Server - Where AI meets revolutionary browser automation! πŸ€–πŸš€

PyDoll MCP Server is a next-generation solution that combines AI with revolutionary browser automation, providing a powerful tool that makes complex web automation tasks simple and efficient.

⚠️ **GitHub.com Fallback** ⚠️