MCP Directory

Playwright MCP (ExecuteAutomation)

Popular community Playwright + API testing MCP server with codegen, screenshots, and device emulation.

Verified
stdio (local)
No auth
TypeScript

Add to your client

Copy the config for your MCP client and paste it into its config file.

Install / run
npx -y @executeautomation/playwright-mcp-server

Paste into ~/Library/Application Support/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "playwright-mcp-executeautomation": {
      "command": "npx",
      "args": [
        "-y",
        "@executeautomation/playwright-mcp-server"
      ]
    }
  }
}

Step-by-step guides: Add to Claude Desktop · Add to Cursor · Add to Windsurf

Before you start

  • Node.js 18+ (run via npx or a global install)
  • Playwright browsers — installed automatically on first use, or manually with `npx playwright install`
  • No API key or authentication required (local stdio server)

About Playwright MCP (ExecuteAutomation)

ExecuteAutomation's Playwright MCP server is a popular community-built bridge that lets an LLM drive a real browser through Microsoft Playwright. It covers the full automation loop — navigate, click, fill, hover, select, screenshot, scrape, and run arbitrary JavaScript — plus device emulation across 143 real device presets (iPhone, iPad, Pixel, Galaxy, desktop).

Beyond UI automation it doubles as an API-testing client: dedicated tools issue HTTP GET/POST/PUT/PATCH/DELETE requests so an agent can exercise REST endpoints in the same session it drives the page. It also exposes Playwright's code-generation workflow, letting you record interactions and emit runnable test code.

It is one of the most widely installed browser MCP servers (available via npm, Smithery, mcp-get, and one-click VS Code install) and runs over stdio with no authentication, making it a quick drop-in for Claude Desktop, Cursor, Cline, and VS Code. Note this is a community project from ExecuteAutomation, distinct from Microsoft's official @playwright/mcp.

Tools & capabilities (12)

playwright_navigate

Open a URL in a Playwright-controlled browser, with optional headless/viewport options

playwright_screenshot

Capture a full-page or element screenshot of the current page

playwright_click

Click an element by selector

playwright_iframe_click

Click an element located inside an iframe

playwright_fill

Type text into an input field

playwright_select

Select an option in a dropdown by value

playwright_hover

Hover the mouse over an element

playwright_evaluate

Execute arbitrary JavaScript in the page context

playwright_console_logs

Retrieve console output (logs, warnings, errors) from the page

playwright_get / post / put / patch / delete

API testing: send HTTP requests against REST endpoints

playwright_close

Close the browser and free resources

playwright codegen

Start/stop a recording session and generate runnable Playwright test code from the captured steps

When to use it

  • Use it when you want an agent to navigate a site, fill forms, and click through a real browser flow
  • Use it when you need screenshots of a page or specific element for visual verification or scraping
  • Use it when you want to record interactions and auto-generate Playwright test code
  • Use it when you need to emulate a specific mobile/tablet device to check responsive behavior
  • Use it when you want to combine UI automation and REST API testing in one session
  • Use it when an agent must scrape or extract content that requires JavaScript execution

Quick setup

  1. 1Ensure Node.js 18+ is installed
  2. 2Add the server to your MCP client config with command `npx -y @executeautomation/playwright-mcp-server` (stdio)
  3. 3Alternatively install globally via `npm install -g @executeautomation/playwright-mcp-server` or one-click VS Code install
  4. 4Restart the MCP client so it picks up the new server
  5. 5Run `npx playwright install` if browsers were not auto-installed on first use
  6. 6Verify by asking the agent to navigate to a URL and take a screenshot

Security notes

Can execute arbitrary JavaScript in pages and generate/run test code, so treat target URLs and any generated scripts as untrusted. The browser shares whatever profile it launches with, so avoid pointing it at sites holding sensitive logged-in sessions.

Playwright MCP (ExecuteAutomation) FAQ

Is this the official Microsoft Playwright MCP server?

No. This is a community server maintained by ExecuteAutomation. Microsoft publishes a separate official server, @playwright/mcp, which has a different toolset (accessibility-tree driven). They are not the same project.

Does it need an API key?

No. It runs locally over stdio and requires no authentication. It controls browsers on your own machine.

Why does it fail to launch a browser on first run?

Playwright's browser binaries may not be installed. Run `npx playwright install` to download Chromium/Firefox/WebKit, then retry.

Can it test REST APIs as well as web pages?

Yes. It exposes dedicated GET/POST/PUT/PATCH/DELETE tools so an agent can exercise API endpoints alongside browser automation in the same session.

Which MCP clients work with it?

Any stdio-compatible client — Claude Desktop, Cursor, Cline, and VS Code (with a one-click install badge). An HTTP transport mode is also available for remote/VS Code setups.

Alternatives to Playwright MCP (ExecuteAutomation)

Compare all alternatives →

All-in-one web access MCP — Web Unlocker, SERP, Scraper API, and a cloud Scraping Browser.

Verified
stdio (local)
API key
JavaScript
12 tools
Updated 14 days agoRepo

Official Browserbase cloud-browser MCP built on Stagehand — natural-language act/extract/observe.

Verified
stdio (local)
API key
TypeScript
8 tools
Updated 14 days agoRepo

Selenium WebDriver MCP server — cross-browser automation for Chrome, Firefox, Edge, and Safari.

Verified
stdio (local)
No auth
JavaScript
18 tools
Updated 3 months agoRepo