Chrome MCP Server transforms your existing Chrome browser into an AI-controlled automation tool, enabling complex tasks like content analysis, semantic search, and interactive browser operations. It targets AI developers and power users seeking to leverage their familiar browser environment for advanced automation without the overhead of separate browser instances.

How It Works

This project utilizes a Chrome extension to act as a Model Context Protocol (MCP) server, exposing browser functionalities via a streamable HTTP connection. This approach allows any AI assistant or agent to interact with the user's current Chrome session, preserving login states and configurations. It offers direct access to Chrome's native APIs, providing a richer feature set compared to Playwright-based solutions.

Quick Start & Requirements

• Installation: Install mcp-chrome-bridge globally via npm or pnpm. Load the downloaded Chrome extension from GitHub releases into Chrome's developer mode.
• Prerequisites: Node.js >= 18.19.0.
• Setup: Requires downloading the extension and installing the bridge package.
• Documentation: English, 中文

Highlighted Details

• Leverages existing Chrome browser, preserving user sessions and configurations.
• Provides full access to Chrome native APIs, including network monitoring and content analysis.
• Features built-in semantic search with a vector database for tab content discovery.
• Includes SIMD-accelerated AI for faster vector operations.
• Offers over 20 tools for browser management, interaction, and data retrieval.

Maintenance & Community

The project is under intensive development, with a roadmap including authentication, recording/playback, and workflow automation. Contributions are welcomed.

Licensing & Compatibility

Licensed under the MIT License, permitting commercial use and integration with closed-source applications.

Limitations & Caveats

The project is in its early stages and is under intensive development, indicating potential for instability and ongoing changes. Firefox extension support is planned but not yet available.