Home / MCP / Browser Automation MCP Server

Browser Automation MCP Server

Runs a local MCP server to drive browser actions via Playwright from an external client.

javascript

11stars

View on GitHub

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
    "mcpServers": {
        "browser_mcp": {
            "command": "node",
            "args": [
                "/path/to/mcp-browser-automation/dist/index.js"
            ]
        }
    }
}

You can run a local MCP server that uses Playwright to automate browser tasks, controlled by an MCP client. This enables you to open a browser, navigate to URLs, and perform scripted actions from an external tool or conversation flow.

How to use

You connect an MCP client to the local MCP server to start and control a browser session. Once connected, you can tell the server to open a browser, load a specific URL, and perform actions such as clicking elements, filling forms, and navigating through pages. Use the client’s commands to start a new browser session, navigate to your target URL, and issue action prompts to the server to drive Playwright-based automation.

How to install

Prerequisites you need installed on your system are Node.js, Playwright, and Claude Desktop.

Install Node.js, then install the project dependencies and build the distribution. Run these commands in your terminal.

npm install

Clone the project, install dependencies, and verify the build artifacts exist in the dist folder.

git clone https://github.com/hrmeetsingh/mcp-browser-automation.git
npm install
ls dist | grep -i index.js || echo "Build artifacts not found in dist/"

Configure the MCP server in Claude Desktop by adding a config entry that points to the local runtime. Create or edit the following JSON configuration in your Claude desktop config.

{
  "mcpServers": {
    "mcp-browser-automation": {
      "command": "node",
      "args": ["/path/to/mcp-browser-automation/dist/index.js"]
    }
  }
}

Additional notes

Start Claude Desktop after configuring the MCP server so you can initiate a browser automation session from a new conversation. This enables you to open a browser, navigate to a URL, and perform actions via Playwright from your MCP client.