Algonius Browser

Algonius Browser

An open-source MCP server that provides browser automation capabilities to external AI systems, enabling navigation, DOM interaction, and web content extraction.

Category
访问服务器

README

<div align="center"> <img src="https://github.com/user-attachments/assets/1b2b1bc0-c7b4-4a45-83f5-4a6161831535" width="600" alt="Algonius Browser Banner" /> </div>

<h1 align="center">MCP Browser Automation</h1>

🌐 Overview

Algonius Browser is an open-source MCP (Model Context Protocol) server that provides browser automation capabilities to external AI systems. It exposes a comprehensive set of browser control tools through the MCP protocol, enabling AI assistants and other tools to navigate websites, interact with DOM elements, and extract web content programmatically.

🎯 Key Features

  • MCP Protocol Integration: Standard interface for AI systems to control browser automation
  • Chrome Extension: Background service worker that handles browser interactions
  • Native Messaging: Go-based MCP host that bridges Chrome extension with external tools
  • Comprehensive Tool Set: 5 browser automation tools + 2 MCP resources
  • Type Safety: Full TypeScript implementation with structured error handling
  • Testing Suite: Comprehensive integration tests for all functionality

🛠️ Available MCP Tools

Navigation & Tabs

  • navigate_to: Navigate to URLs with configurable timeout handling
  • manage_tabs: Create, close, and switch between browser tabs

DOM Interaction

  • get_dom_extra_elements: Advanced DOM element extraction with pagination and filtering
  • click_element: Click DOM elements using CSS selectors or text matching
  • set_value: Set values in input fields, textareas, and form elements
  • scroll_page: Scroll pages up or down with customizable distances

📋 Available MCP Resources

Browser State Resources

  • browser://current/state: Complete current browser state in AI-friendly Markdown format

    • Active tab information
    • All browser tabs with URLs, titles, and status
    • Real-time state updates via resource notifications
  • browser://dom/state: Current DOM state overview in Markdown format

    • Page metadata (URL, title, scroll position)
    • First 20 interactive elements
    • Total element count with "more available" indicators
    • Simplified DOM structure
    • Auto-updates when page changes

🚀 Quick Start

1. Install Chrome Extension

From Chrome Web Store (Recommended):

<a href="https://chromewebstore.google.com/detail/algonius-browser-mcp/fmcmnpejjhphnfdaegmdmahkgaccghem" target="_blank"> <img src="https://github.com/user-attachments/assets/4c2c0b5e-8f63-4a8b-9a5e-2d7e8f3c9b1a" alt="Available in the Chrome Web Store" width="248" height="75"> </a>

  1. Click the "Add to Chrome" button on the Chrome Web Store page
  2. Confirm the installation when prompted
  3. The extension will be automatically installed and ready to use

From Source (Development):

# Clone and build
git clone https://github.com/algonius/algonius-browser.git
cd algonius-browser
pnpm install
pnpm build

# Load in Chrome
# 1. Open chrome://extensions/
# 2. Enable "Developer mode"
# 3. Click "Load unpacked"
# 4. Select the 'dist' folder

⚠️ Important: The Chrome extension requires the MCP Host backend service to function properly. Please continue with step 2 to complete the installation.

2. Install MCP Host

One-Click Installation (Recommended):

Linux/macOS:

curl -fsSL https://raw.githubusercontent.com/algonius/algonius-browser/master/install-mcp-host.sh | bash

Windows (PowerShell):

iwr -useb https://raw.githubusercontent.com/algonius/algonius-browser/master/install-mcp-host.ps1 | iex

Manual Installation:

# Download latest release
wget https://github.com/algonius/algonius-browser/releases/latest/download/mcp-host-linux-x86_64.tar.gz

# Extract and install
tar -xzf mcp-host-linux-x86_64.tar.gz
cd mcp-host-linux-x86_64
./install.sh

3. Verify Installation

# Test the MCP host installation
mcp-host-go --version

# The MCP host will be automatically started when needed by the Chrome extension
# You should see the extension icon in your Chrome toolbar

Success: Both components are now installed! The Chrome extension will automatically communicate with the MCP Host when browser automation is requested.

🔧 Integration Examples

Using with AI Assistants

Once installed, AI systems can use the browser automation tools and resources through the MCP protocol:

Tool Usage:

{
  "method": "tools/call",
  "params": {
    "name": "navigate_to",
    "arguments": {
      "url": "https://example.com",
      "timeout": 30000
    }
  }
}

Resource Access:

{
  "method": "resources/read",
  "params": {
    "uri": "browser://current/state"
  }
}

Common Workflows

Web Scraping:

  1. navigate_to → Navigate to target site
  2. Read browser://dom/state → Get page overview
  3. get_dom_extra_elements → Get specific elements with pagination
  4. click_element → Interact with elements
  5. Read browser://dom/state → Extract updated content

Form Automation:

  1. navigate_to → Go to form page
  2. Read browser://dom/state → Identify form elements
  3. set_value → Fill form fields
  4. click_element → Submit form
  5. Read browser://current/state → Verify completion

Multi-Tab Management:

  1. Read browser://current/state → Check current tabs
  2. manage_tabs → Create/switch tabs
  3. navigate_to → Load content in each tab
  4. Read browser://current/state → Monitor all tab states

Page Navigation with Scrolling:

  1. navigate_to → Go to target page
  2. Read browser://dom/state → Get initial page state
  3. scroll_page → Scroll to load more content
  4. get_dom_extra_elements → Extract newly loaded elements

🏗️ Architecture

External AI System
       ↓ (MCP Protocol)
   MCP Host (Go)
       ↓ (Native Messaging)
Chrome Extension
       ↓ (Chrome APIs)
    Browser Tabs

Components

  • MCP Host: Go-based native messaging host that implements MCP protocol
  • Chrome Extension: Background service worker with tool handlers
  • Content Scripts: DOM interaction and data extraction utilities
  • Integration Tests: Comprehensive test suite for all tools

🧪 Development

Build from Source

Prerequisites:

  • Node.js 22.12.0+
  • pnpm 9.15.1+
  • Go 1.21+ (for MCP host)

Build Extension:

pnpm install
pnpm build

Build MCP Host:

cd mcp-host-go
make build

Run Tests:

# Extension tests
pnpm test

# MCP host tests  
cd mcp-host-go
make test

Development Mode

# Extension development
pnpm dev

# MCP host development
cd mcp-host-go
make dev

📊 Supported Platforms

MCP Host:

  • Linux x86_64
  • macOS Intel (x86_64) and Apple Silicon (arm64)
  • Windows x86_64

Chrome Extension:

  • Chrome/Chromium 88+
  • Microsoft Edge 88+

📚 Documentation

Detailed documentation available in the docs/ directory:

🤝 Contributing

We welcome contributions! Check out our CONTRIBUTING.md for guidelines.

Ways to contribute:

  • Report bugs and feature requests
  • Submit pull requests for improvements
  • Add integration tests
  • Improve documentation
  • Share usage examples

🔒 Security

For security vulnerabilities, please create a GitHub Security Advisory rather than opening a public issue.

💬 Community

📄 License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

👏 Acknowledgments

Built with these excellent open-source projects:


Made with ❤️ by the Algonius Browser Team

Give us a star 🌟 if this project helps you build better browser automation!

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选