MCP Playwright Server

MCP Playwright Server is an MCP server that provides browser automation capabilities using the Playwright framework. It connects Model Context Protocol clients like Claude Desktop directly to headless or headed browser engines including Chromium, Firefox, and WebKit. Built for QA engineers, software developers, and automated workflow designers, it enables AI assistants to execute structured web operations through standard model function calls. Users can launch dedicated browser instances, open new tabs or pages, navigate to target URLs, interact with web elements via clicks and text input, and extract plain text from document nodes. The server also supports taking visual screenshots for interface verification, waiting for dynamic DOM selectors, and managing browser lifecycle events like closing pages or instances. Running locally or inside a preconfigured Docker container, it allows external AI agents to perform automated regression testing, verify web application state, and gather data from complex single-page applications without manual intervention.

Category: Browser & Web Automation

Tags: automation, headless browser, playwright, testing, web scraping

Visit MCP Playwright Server

How to install and configure MCP Playwright Server

1. Install the global package using npm: bash npm install -g mcp-playwright Alternatively, clone the repository, run npm install, and execute npm run install:browsers to set up Playwright binaries. 2. Open your Claude Desktop configuration file (such as ~/Library/Application Support/Claude/claude_desktop_config.json on macOS). 3. Register the server under mcpServers using Node.js: json { "mcpServers": { "playwright": { "command": "node", "args": ["/path/to/your/mcpplaywright/src/index.js"], "env": { "PLAYWRIGHT_HEADLESS": "true" } } } } Or configure it to run via Docker using "command": "docker" with "args": ["run", "--rm", "-i", "mcp-playwright"]. 4. Restart Claude Desktop to activate the Playwright tools.

What you can do with MCP Playwright Server

• Automate end-to-end testing of user authentication flows and verify error messages upon entering invalid credentials. * Extract text content and structured tabular metrics from dynamic single-page web applications for reporting. * Capture full-page screenshots to perform visual regression checks on newly deployed application interfaces. * Fill and submit multi-field online forms by targeting specific input elements and automated button clicks. * Monitor dynamic user interfaces by configuring selector waits to inspect element visibility and rendered text state.

Key facts

• Open Source
• https://github.com/MostafaRaafat91/MCP-Playwright-Server
• Browser & Web Automation, Developer Tools & Code Intelligence
• automation, headless browser, playwright, testing, web scraping

Part of MCP Servers

How do I install MCP Playwright Server?

You can install it globally via npm using the command npm install -g mcp-playwright and start it with mcp-playwright start. Alternatively, you can clone the repository, install dependencies with npm install, fetch browser dependencies via npm run install:browsers, or build and run the provided Docker image using npm run docker:build and npm run docker:run.

What can MCP Playwright Server do?

MCP Playwright Server provides tools to control Chromium, Firefox, and WebKit browsers. It allows AI clients to launch browsers, create new pages, navigate to URLs, click elements, type text into inputs, retrieve rendered text, capture screenshots, wait for DOM selectors, and terminate browser processes.

Which MCP clients work with MCP Playwright Server?

The server works with Claude Desktop and any client supporting the Model Context Protocol. Configuration can be completed by updating the MCP settings file to run the server script directly via Node or through an interactive Docker container.

Is MCP Playwright Server open source?

Yes, MCP Playwright Server is open source and distributed under the MIT license. You can review the source code, contribute enhancements, and inspect implementation details directly on its GitHub repository.

Can I run browsers in headed mode with MCP Playwright Server?

Yes, you can control browser visibility by setting the PLAYWRIGHT_HEADLESS environment variable to false in your MCP configuration. You can also configure the PLAYWRIGHT_SLOWMO environment variable to add a delay in milliseconds between automated actions.