# MCP Playwright

> A Model Context Protocol server that provides browser automation capabilities using Playwright for LLM interactions.

MCP Playwright is a Model Context Protocol (MCP) server that enables browser automation capabilities using Playwright, allowing Large Language Models (LLMs) to interact with web pages, take screenshots, generate test code, and execute JavaScript in a real browser environment.

This tool bridges the gap between AI assistants and browser automation, making it possible for LLMs to perform complex web interactions through a standardized protocol. It supports multiple transport mechanisms and integrates seamlessly with popular AI development environments.

- **Browser Automation** - Navigate to URLs, click elements, fill forms, select options, hover over elements, and perform comprehensive web interactions through Playwright's powerful automation engine.

- **Screenshot Capabilities** - Capture full-page or element-specific screenshots, with support for console log retrieval to help debug and analyze web page behavior.

- **Test Code Generation** - Automatically generate Playwright test code based on interactions, streamlining the test automation workflow for developers.

- **JavaScript Execution** - Execute custom JavaScript code in the browser console, enabling advanced page manipulation and data extraction.

- **API Testing Support** - Perform GET and POST requests directly through the MCP server, allowing for comprehensive API testing alongside browser automation.

- **Multiple Transport Options** - Supports both SSE (Server-Sent Events) and stdio transport mechanisms for flexible integration with different client applications.

- **Vision Mode** - Includes accessibility snapshot and vision-based interaction modes for enhanced element detection and interaction capabilities.

- **Multi-Client Support** - Works with Claude Desktop, VS Code with GitHub Copilot, Cursor IDE, and other MCP-compatible clients.

To get started, install the package via npm with `npx @anthropic-ai/create-mcp-server` or configure it directly in your MCP client settings. The server can be run in SSE mode for web-based clients or stdio mode for desktop applications. Configuration involves adding the server to your client's MCP settings file with the appropriate command and arguments.

## Features
- Browser navigation and interaction
- Screenshot capture (full page and element)
- Form filling and submission
- Element clicking and hovering
- Dropdown selection
- JavaScript execution in browser
- Playwright test code generation
- API testing (GET and POST requests)
- Console log retrieval
- Accessibility snapshot mode
- Vision-based interaction mode
- SSE and stdio transport support
- Multi-tab management

## Integrations
Claude Desktop, VS Code, GitHub Copilot, Cursor IDE, Playwright

## Platforms
WINDOWS, MACOS, LINUX

## Pricing
Open Source

## Links
- Website: https://executeautomation.github.io/mcp-playwright/
- Documentation: https://executeautomation.github.io/mcp-playwright/
- Repository: https://github.com/executeautomation/mcp-playwright
- EveryDev.ai: https://www.everydev.ai/tools/mcp-playwright