Chrome

You can run and control a Chrome browser automation server from MCP clients to perform web automation with AI models. This server lets you navigate pages, interact with elements, extract data, manage browser state, and access system features through a standard MCP interface.

How to use

You use this server by connecting an MCP client to the server over WebSocket. Start the server on your machine, then configure your MCP client to point to the server’s WebSocket URL. Once connected, you can send commands to navigate to URLs, click elements, type text, execute JavaScript, and extract data from pages. Always include a valid tab identifier for each operation and verify the tab is ready before performing actions.

How to install

Prerequisites include a Python 3.12 runtime and a Chrome browser extension that acts as a WebSocket client. You also need a runtime method to start the MCP server, such as uv or Docker.

If you are using uv to run the server locally, choose the appropriate setup for your operating system and insert the exact command and environment variables shown in the setup blocks below.

{
  "mcpServers": {
    "mcp-chrome-integration-windows": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "%LOCALAPPDATA%\\Programs\\mcp-chrome-integration\\src",
        "mcp-server"
      ],
      "env": {
        "WEBSOCKET_PORT": "8012"
      }
    }
  }
}

Additional setup for Linux and macOS

{
  "mcpServers": {
    "mcp-chrome-integration": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/usr/local/bin/mcp-chrome-integration/src",
        "mcp-server"
      ],
      "env": {
        "WEBSOCKET_PORT": "8012"
      }
    }
  }
}

Additional setup for other OSes

{
  "mcpServers": {
    "mcp-chrome-integration": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/usr/local/bin/mcp-chrome-integration/src",
        "mcp-server"
      ],
      "env": {
        "WEBSOCKET_PORT": "8012"
      }
    }
  }
}

Available tools

tool_navigate_to

Navigate the browser to a specified URL in a given tab.

tool_click_element

Click a page element identified by a selector within a specific tab.

tool_type_text

Type text into an input field identified by a selector in a given tab.

tool_state

Check the current state of a tab, including loading status and visibility of elements.

tool_execute_script

Execute a JavaScript snippet in the context of the current page and return results.

tool_extract_table

Extract data from a table element matching a selector and return structured rows.

tool_get_element_info

Retrieve information about an element, such as dimensions and visibility.