Selenium MCP Server

A Model Context Protocol (MCP) server that provides browser automation capabilities using Selenium WebDriver.

Features

• Browser Automation: Navigate, click, type, and interact with web pages

• Element Interaction: Find elements by CSS selectors, wait for elements, get text content

• Screenshots: Capture screenshots of the current page

• JavaScript Execution: Execute custom JavaScript in the browser

• Page Information: Get page title and current URL

• Headless Mode: Run browser in headless mode for CI/CD environments

Related MCP server: mcp-chrome-server

Prerequisites

• Node.js (v16 or higher)

• Chrome browser (for WebDriver)

• npm

Installation

1. Install dependencies:

npm install

2. Configure the server by editing config.json:

{
  "selenium": {
    "browser": "chrome",
    "headless": true,
    "timeout": 30000,
    "windowSize": {
      "width": 1920,
      "height": 1080
    }
  }
}

Usage

Starting the Server

npm start

The server runs on stdio and communicates using the MCP protocol.

Available Tools

The server provides the following tools:

1. navigate - Navigate to a URL

   • Parameters: url (string, required)

2. click - Click on an element

   • Parameters: selector (string, required) - CSS selector

3. type - Type text into an element

   • Parameters: selector (string, required), text (string, required)

4. get_text - Get text content of an element

   • Parameters: selector (string, required)

5. wait_for_element - Wait for an element to be present

   • Parameters: selector (string, required), timeout (number, optional)

6. screenshot - Take a screenshot

   • Parameters: filename (string, optional)

7. get_title - Get the page title

   • Parameters: none

8. get_url - Get the current URL

   • Parameters: none

9. execute_script - Execute JavaScript

   • Parameters: script (string, required)

10. close_browser - Close the browser session

    • Parameters: none

Configuration

Edit config.json to customize:

• browser: Browser to use (chrome, firefox, etc.)

• headless: Run in headless mode (true/false)

• timeout: Default timeout for operations (milliseconds)

• windowSize: Browser window dimensions

Example Usage

Here's an example of how to use the server tools:

1. Navigate to a page:

navigate: { "url": "https://example.com" }

2. Wait for an element:

wait_for_element: { "selector": "#login-button" }

3. Type into an input:

type: { "selector": "#username", "text": "user123" }

4. Click a button:

click: { "selector": "#login-button" }

5. Take a screenshot:

screenshot: { "filename": "login.png" }

6. Close the browser:

close_browser: {}

Troubleshooting

• Ensure Chrome browser is installed

• Make sure ChromeDriver is compatible with your Chrome version

• Check that the port in config.json is not already in use

• Verify Node.js version is 16 or higher

License

ISC

AI_Automation_Framework_With_MCP_Server

This is an Automation framework which is built with AI using MCP Server,Java MCP Client and Automated Flows of Google Search