Selenium MCP Server

Example Use Cases

• 1

  Navigate and scrape dynamic websites, interact with forms, and capture full-page screenshots via a controllable browser.

• 2

  Execute custom JavaScript in the page context and capture both return values and console output to extract structured data or debug scripts.

• 3

  Monitor browser console and network logs, manage localStorage, and interact with elements inside iframes using XPath or attribute-based selectors.

Description

This project implements a Model Context Protocol (MCP) server wrapping Selenium WebDriver to provide web automation capabilities for AI assistants. It offers navigation, element discovery and interaction, screenshots, style inspection, JavaScript execution (with console capture), browser and network logs, and localStorage management, with iframe and XPath support. The server connects to Chrome via the remote debugging protocol (or can auto-start Chrome) and communicates over stdio to integrate with MCP-compatible clients like Claude Desktop or VS Code Copilot.