Playwright
Playwright MCP by Microsoft gives AI agents deterministic browser automation driven by structured accessibility snapshots instead of screenshots. It can navigate pages, click, type, fill forms, capture snapshots, intercept network traffic, and run end-to-end tests across Chromium, Firefox, and WebKit.
Install Playwright
- Published by
- Official (vendor)
- Transport
- Local process (stdio)
- Authentication
- None
- Package
- @playwright/mcp
Claude Desktop, Claude Code and Cursor
Add this to the mcpServers object in your client's
config file, then restart the client.
{
"mcpServers": {
"playwright": {
"command": "npx",
"args": [
"@playwright/mcp@latest"
]
}
}
}VS Code uses a different key — show that config
Identical entry, filed under servers rather than
mcpServers. Put it in
.vscode/mcp.json for one workspace.
{
"servers": {
"playwright": {
"command": "npx",
"args": [
"@playwright/mcp@latest"
]
}
}
}Before it will answer
Needs Node.js 18 or newer and nothing else. It drives a real browser, so treat any page it visits as untrusted input — a web page can carry instructions aimed at the agent reading it.
Tools
browser_click
Perform a click on an element identified in the page snapshot.
browser_close
Close the current page and release its browser resources.
browser_console_messages
Return all console messages logged by the current page.
browser_drag
Perform a drag-and-drop between two elements on the page.
browser_evaluate
Evaluate a JavaScript expression on the page or a chosen element.
browser_file_upload
Upload one or multiple files through a file input element.
browser_fill_form
Fill multiple form fields in a single structured action.
browser_find
Search the accessibility snapshot for matching text or a regex.
browser_generate_locator
Generate a stable Playwright locator for a chosen element.
browser_handle_dialog
Accept or dismiss a native browser dialog such as an alert.
browser_hover
Hover the pointer over an element to reveal menus or tooltips.
browser_navigate
Navigate the browser to a specific URL and load the page.
browser_navigate_back
Go back to the previous page in the browser history.
browser_network_requests
List the network requests issued since loading the page.
browser_pdf_save
Save the current page as a PDF document to a file on disk.
browser_press_key
Press a keyboard key such as Enter, Tab, or an arrow key.
browser_resize
Resize the browser window to the given width and height.
browser_select_option
Select one or more options in a dropdown or select element.
browser_snapshot
Capture the accessibility snapshot of the current page for the agent.
browser_tabs
List, create, close, or select a browser tab in the session.
browser_take_screenshot
Take a screenshot of the current page or a specific element.
browser_type
Type text into an editable element such as an input or textarea.
browser_verify_element_visible
Assert that a given element is visible on the page.
browser_wait_for
Wait for text to appear or disappear, or for a timeout to elapse.