Playwright MCP Server

Example Use Cases

• 1

  Open and control a Chromium browser to interact with web pages (click, navigate, open/close tabs) programmatically.

• 2

  Extract readable page content (readability) from the active tab for downstream NLP or summarization tasks.

• 3

  A simple MCP interface to orchestrate multi-tab browsing workflows and manage browser lifecycle from AI controllers.

Description

This project implements an MCP server that manipulates a Chromium browser using Playwright, exposing simple RPC-style tools (open-browser, open-new-page, navigate-to-page, close-page, close-browser, get-page-content). It's intended to enable AI-driven browser control and page content retrieval (readability) for agent workflows. The repo requires bun and pnpm to install and provides example MCP configuration and environment variable settings for Chrome paths and user data. It's presented as an experimental/community tool rather than a commercial product.