Browser Automation MCP Server

A Model Context Protocol (MCP) server that provides browser automation capabilities for Claude and other MCP-compatible AI assistants.

Features

• Web Automation: Control web browsers programmatically
• Screenshot Capture: Take screenshots of web pages
• Element Interaction: Click, type, and interact with web elements
• Navigation: Navigate between pages and manage browser state
• Form Filling: Automate form filling and submission
• Data Extraction: Extract data from web pages

Installation

# Clone the repository
git clone https://github.com/samihalawa/browser-automation-server.git
cd browser-automation-server

# Install dependencies
npm install

# Build the server
npm run build

Usage

Starting the Server

npm start

Configuration

Add the server to your MCP configuration:

{
  "servers": {
    "browser-automation": {
      "command": "/path/to/node",
      "args": ["/path/to/browser-automation-server/build/index.js"],
      "enabled": true,
      "port": 3008,
      "environment": {
        "NODE_PATH": "/path/to/node_modules",
        "PATH": "/usr/local/bin:/usr/bin:/bin"
      }
    }
  }
}

Available Tools

navigate

Navigate to a URL.

Parameters:

• url (string, required): URL to navigate to
• waitUntil (string, optional): Navigation wait condition. Options: 'load', 'domcontentloaded', 'networkidle'. Default: 'load'

screenshot

Take a screenshot of the current page.

Parameters:

• fullPage (boolean, optional): Whether to capture full page or just viewport. Default: false
• path (string, optional): Path to save the screenshot to. If not provided, returns base64 encoded image

click

Click on an element.

Parameters:

• selector (string, required): CSS selector of the element to click
• waitForSelector (boolean, optional): Whether to wait for the selector to appear. Default: true

type

Type text into an input field.

Parameters:

• selector (string, required): CSS selector of the input element
• text (string, required): Text to type
• delay (number, optional): Delay between keystrokes in milliseconds. Default: 0

extract

Extract data from the page.

Parameters:

• selector (string, required): CSS selector of the elements to extract
• attribute (string, optional): Attribute to extract. If not provided, extracts text content

evaluate

Evaluate JavaScript in the browser context.

Parameters:

• script (string, required): JavaScript code to evaluate
• args (array, optional): Arguments to pass to the script

Example Usage

1. Navigate to a website:

   navigate(url: "https://example.com")
   

2. Take a screenshot:

   screenshot(fullPage: true)
   

3. Click a button:

   click(selector: "#submit-button")
   

4. Fill a form:

   type(selector: "#username", text: "user123")
   type(selector: "#password", text: "password123")
   click(selector: "#login-button")
   

5. Extract data:

   extract(selector: ".product-title", attribute: "innerText")
   

Requirements

• Node.js 14+
• Playwright for browser automation

License

MIT