Selenium MCP Server

A powerful VS Code extension to control and automate browsers using your Selenium MCP server.

Features

• Start/stop the Selenium MCP server from VS Code
• Status bar indicator for server status
• Manage browser sessions (Chrome/Firefox, headless/full) in a sidebar
• Start/stop browser sessions
• Navigate, click, type, and take screenshots in browser sessions
• View server logs and screenshots directly in VS Code
• Configurable Python path, arguments, and working directory
• Quick Start onboarding and helpful error messages
• Full MCP tool coverage: All Selenium MCP server tools are available as Copilot/LLM tools

Available Tools

All of the following tools are registered and callable from Copilot, LLMs, and the MCP protocol. Usage examples show how to call each tool with sample arguments:

• start_browser: Launches browser with enhanced session management

  {
    "name": "start_browser",
    "arguments": { "browser": "chrome", "options": { "headless": true } }
  }
  

• list_sessions: Lists all active browser sessions

  { "name": "list_sessions", "arguments": {} }
  

• switch_session: Switches to a different browser session

  { "name": "switch_session", "arguments": { "session_id": "your-session-id" } }
  

• close_session: Closes a specific browser session

  { "name": "close_session", "arguments": { "session_id": "your-session-id" } }
  

• navigate: Navigates to a URL with enhanced error handling

  { "name": "navigate", "arguments": { "url": "https://example.com", "wait_for_load": true } }
  

• find_element: Finds an element with enhanced waiting

  { "name": "find_element", "arguments": { "by": "css", "value": "#my-element", "timeout": 5000 } }
  

• click_element: Clicks an element with enhanced error handling

  { "name": "click_element", "arguments": { "by": "css", "value": "#my-button", "force_click": true } }
  

• send_keys: Sends keys to an element with enhanced typing

  { "name": "send_keys", "arguments": { "by": "css", "value": "#input", "text": "hello", "clear_first": true } }
  

• get_element_text: Gets the text of an element

  { "name": "get_element_text", "arguments": { "by": "css", "value": "#output" } }
  

• hover: Moves the mouse to hover over an element

  { "name": "hover", "arguments": { "by": "css", "value": "#hover-target" } }
  

• drag_and_drop: Drags an element and drops it onto another element

  { "name": "drag_and_drop", "arguments": { "by": "css", "value": "#source", "targetBy": "css", "targetValue": "#target" } }
  

• double_click: Performs a double click on an element

  { "name": "double_click", "arguments": { "by": "css", "value": "#element" } }
  

• right_click: Performs a right click (context click) on an element

  { "name": "right_click", "arguments": { "by": "css", "value": "#element" } }
  

• press_key: Simulates pressing a keyboard key

  { "name": "press_key", "arguments": { "key": "Enter" } }
  

• upload_file: Uploads a file using a file input element

  { "name": "upload_file", "arguments": { "by": "css", "value": "input[type='file']", "filePath": "C:/Users/YourName/file.txt" } }
  

• take_screenshot: Captures a screenshot of the current page

  { "name": "take_screenshot", "arguments": { "full_page": true } }
  

• wait_for_element: Waits for an element to be present and optionally visible

  { "name": "wait_for_element", "arguments": { "by": "css", "value": "#wait-for-me", "wait_for_visible": true } }
  

• execute_script: Executes JavaScript code in the browser

  { "name": "execute_script", "arguments": { "script": "return document.title;" } }
  

• get_page_info: Gets information about the current page

  { "name": "get_page_info", "arguments": { "include_title": true, "include_url": true, "include_source": false } }
  

• get_server_version: Returns the current version of the Selenium MCP server

  { "name": "get_server_version", "arguments": {} }
  

Each tool exposes all its parameters and is fully documented in the extension and MCP server docs.

Requirements

• Python 3.7+
• Selenium MCP Server (pip install selenium-mcp-server)
• Chrome or Firefox browser installed

Quick Start

1. Install the Selenium MCP server (see server repo)
2. Install this extension in VS Code (search for "Selenium MCP Server" or krishnapollu.selenium-mcp-server)
3. Start the MCP server: Use the status bar or Command Palette (Start Selenium MCP Server)
4. Open the MCP Sessions sidebar: Manage browser sessions
5. Right-click a session for automation actions (navigate, click, type, screenshot, logs)

Installation

Selenium MCP Server on Visual Studio Marketplace

• From Marketplace:
  Search for Selenium MCP Server in the VS Code Extensions view and click Install.
• From VSIX:
  Download the latest .vsix from Marketplace or your GitHub releases, then run:
  • Open Command Palette → Extensions: Install from VSIX...

Usage

• Start/Stop Server: Use the status bar or Command Palette
• Manage Sessions: Use the MCP Sessions sidebar
• Automation Actions: Right-click a session for:
  • Navigate Session
  • Click Element (by CSS selector)
  • Type Into Element
  • Take Screenshot
  • Show Server Logs
• Refresh Sessions: Use the refresh button in the sidebar
• Show Quick Start: Show MCP Quick Start command in the palette

Prompting Cheat Sheet

Use these prompts and patterns to get the most out of the Selenium MCP Server extension and MCP server:

For AI/LLM Clients (Copilot, Cursor, Claude, etc.)

• Call any tool directly:

  Use the selenium_mcp_server.<tool_name> tool with the required parameters. Example: Call selenium_mcp_server.navigate with { "url": "https://example.com" }.

• Start a session:

  Start a new Chrome session and navigate to https://example.com.

• Interact with elements:

  Click the login button with selector #login. Type 'myusername' into the username field and 'mypassword' into the password field, then click submit.

• Take screenshots or get logs:

  Take a screenshot of the current page and show it in VS Code. Show me the browser logs for this session.

• Session management:

  List all active browser sessions and close the one running in headless mode.

For Manual Use in VS Code

• Use the Command Palette for actions like:
  • Start Selenium MCP Server
  • Start Browser Session
  • Navigate Session
  • Click Element
  • Type Into Element
  • Take Screenshot
  • Show Server Logs
• Right-click sessions in the MCP Sessions sidebar for context actions.

For Custom Scripts (MCP Client API)

• Send JSON commands to the MCP server, for example:

  { "name": "navigate", "arguments": { "session_id": "abc123", "url": "https://example.com" } }
  

• Chain commands for automation flows.

Prompting Tips

• Be specific: Specify browser, session, selector, and action.
• Use step-by-step instructions for complex flows.
• Request feedback: Ask for screenshots, logs, or confirmation after actions.
• Leverage session management: Start, stop, and list sessions as needed.

Configuration

Set these in your VS Code settings:

• selenium-mcp-server.pythonPath: Path to Python executable
• selenium-mcp-server.serverArgs: Arguments for MCP server (default: -m selenium_mcp_server)
• selenium-mcp-server.workingDirectory: Working directory for the server

Troubleshooting

• Server not starting?
  • Check your Python path and MCP server install
  • See Output Channel for errors
• No sessions listed?
  • Make sure the server is running
  • Refresh the sidebar
• Automation commands not working?
  • Check server logs for errors
  • Ensure browser drivers are installed

Onboarding

• On first activation, a Quick Start guide is shown
• Use the Show MCP Quick Start command anytime
• See this README for full details

Contact & Support

• For MCP server issues: selenium-mcp-server repo
• For extension issues: Open an issue in this repo

Contributing

Contributions, issues, and feature requests are welcome! Feel free to check the issues page or submit a pull request.

License

MIT

Enjoy automating browsers with the power of VS Code and Selenium MCP!