Playwright

OFFICIALGitHub

Playwright-mcp gives LLMs access to browser automation.

Most interesting: It uses accessibility features, so the LLM doesn't need to try and parse screenshots.

From the Readme:

Snapshot Mode

  • browser_navigate - Navigate to a URL
  • browser_go_back - Go back to the previous page
  • browser_go_forward - Go forward to the next page
  • browser_click - Perform click on a web page
  • browser_hover - Hover over element on page
  • browser_drag - Perform drag and drop between two elements
  • browser_type - Type text into editable element
  • browser_press_key - Press a key on the keyboard
  • browser_snapshot - Capture accessibility snapshot of the current page (better than screenshot)
  • browser_save_as_pdf - Save page as PDF
  • browser_wait - Wait for a specified time in seconds
  • browser_close - Close the page

Vision Mode

  • browser_navigate - Navigate to a URL
  • browser_go_back - Go back to the previous page
  • browser_go_forward - Go forward to the next page
  • browser_screenshot - Capture screenshot of the current page
  • browser_move_mouse - Move mouse to specified coordinates
  • browser_click - Click at specified coordinates
  • browser_drag - Perform drag and drop operation
  • browser_type - Type text at specified coordinates
  • browser_press_key - Press a key on the keyboard
  • browser_save_as_pdf - Save page as PDF
  • browser_wait - Wait for a specified time in seconds
  • browser_close - Close the page

https://simonwillison.net/2025/Mar/25/playwright-mcp/