Playwright MCP
agent-ready json output non-interactive
Microsoft's official MCP server for browser automation via Playwright. Allows agents to navigate pages, click, type, take screenshots, and execute JavaScript.
How to install Playwright MCP
npx @playwright/mcp@latest When to use Playwright MCP
- Automating multi-step web workflows like filling forms, clicking through paginated lists, or scraping dynamic content that requires JavaScript execution.
- Capturing screenshots or generating PDFs of web pages for reporting, monitoring, or archival purposes, especially when full page or specific element captures are needed.
- Intercepting and inspecting network requests/responses during web application testing or debugging to verify API calls or modify request/response data.
When not to use Playwright MCP
- Performing simple HTTP requests or fetching static HTML where a lightweight tool like `curl` or `fetch` suffices without the overhead of a full browser.
- Scraping data from websites that require no JavaScript execution and have simple, well-structured HTML that can be parsed with traditional HTML parsers.
Playwright MCP features
- · Full browser control
- · MCP-native protocol
- · Screenshot + PDF export
- · Form interaction
- · Network interception
Want your agent to find this automatically?
Add the MCP server to your agent config and it will discover tools like Playwright MCP on its own.
Set up MCP →