The Playwright MCP server, maintained by Microsoft, gives AI models live control of a Chromium browser via the Model Context Protocol. Navigate to any URL, click elements, fill forms, extract text, and take screenshots — all from a natural language conversation.
Bring your own
Claude Sonnet 4.5
Chat with 60+ AI models on the same workflow — switch to a different model mid-conversation and re-run the same prompt, or use Compare mode to put several side-by-side and balance quality vs. cost.
No token required
How models use it and what it is built for.
Playwright MCP wraps Microsoft's battle-tested browser automation library as MCP tools. When an AI model calls browser_navigate, a real Chromium tab opens and loads the URL. Subsequent calls can click any element by selector, type into inputs, submit forms, evaluate JavaScript, or capture a full-page screenshot. The server runs locally on your machine via npx — no account, no API key, no infrastructure. It is the go-to choice for agent-driven end-to-end testing, live web research, form filling and content extraction workflows.
Typical tools an AI model can call. Exact names vary by version.
Copy any of these into MCP Agent Studio after connecting.
Go to news.ycombinator.com and list the top 10 story titles with their point counts.
Navigate to our staging site, log in with these credentials and screenshot the dashboard.
Fill out the contact form at this URL and capture a screenshot of the confirmation page.
Scrape the pricing table from this URL and format it as a clean markdown table.
This is not a single-model product: you get the same MCP connection with 60+ models (Claude, GPT, Gemini, DeepSeek, open-weight, and more), you can switch mid-conversation, and you can open Compare mode to run the same prompt against multiple models at once. The card above is a suggested starting point for this server — not the only choice.
Default pick for Playwright
Claude Sonnet 4.5
Claude Sonnet 4.5 handles multi-step browser workflows and visual reasoning best. Use Haiku for simple scraping; GPT-5.2 for tasks requiring vision on screenshots.
Listing tools proves the server is reachable, not that a model can work with it. Evals go further: they read every tool on the server, write a test suite from its real schemas, and run it — code decides pass/fail on the responses (schema conformance, error codes, pagination, result caps) while a scoring model grades plain-English tasks driven through the tools.
Get a pass/fail report per tool with the evidence behind each verdict — and replay the same suite after every schema change. Destructive tools are excluded from the run.
Run evalsOpen MCP Agent Studio with the connection pre-filled. Add your token, pick any of 60+ models, and start chatting — no install required.
Open Agent StudioCommon questions about connecting, scoping and using it safely.
It is an open-source MCP server from Microsoft that gives AI models live control of a Chromium browser. Instead of reasoning about a screenshot you paste, the agent can navigate, click, type and extract data from any live website in real time.
Playwright downloads and bundles its own Chromium build, so you do not need Chrome or any other browser pre-installed. Run npx @playwright/mcp@latest and Chromium is ready automatically.
Yes. The server is MIT-licensed and completely free. You pay only for AI model usage — or use the free tier in MCP Agent Studio. There is no Playwright subscription or API key.
The standard distribution runs locally (stdio transport). For cloud use, you would need to run the server process on a remote machine and expose it over HTTP/SSE, or use a hosted browser-automation platform that wraps Playwright.
Scraping APIs return raw HTML or pre-extracted data for a fixed schema. Playwright MCP gives the AI agent full interactive browser control — it can log in, click through multi-step flows, react to what it sees, and decide the next action dynamically.