Playwright MCP

Example Use Cases

• 1

  Navigate web pages, open/close/select tabs, and perform deterministic actions like clicking, typing, and selecting options using accessibility snapshots.

• 2

  Extract structured data from pages or fill forms programmatically without relying on screenshots or vision models.

• 3

  Run automated web tests or background/scheduled browser tasks (headless or headed) with programmatic control over browser launch options and transports.

Description

Playwright MCP runs a Playwright-based Model Context Protocol server that lets LLMs interact with web pages using structured accessibility trees (rather than pixel input). It supports snapshot-based actions (click, type, drag, select) for deterministic, LLM-friendly automation and an optional vision mode for screenshot-based interactions. The server offers CLI options for browser selection, headless/headed modes, capabilities, and transports, and includes tab, navigation, file, and utility tools for full browser control.