remorses/playwriter

Chrome extension & CLI to let agents control your browser. Runs Playwright snippets in a stateful sandbox. Available as CLI or MCP

What it solves

Playwriter allows AI agents to control a user's actual, running Chrome browser instead of spawning a fresh, isolated instance. This solves the problem of automation tools being flagged by bot detectors and the need to manually handle logins, cookies, and extensions, as the agent operates within the user's existing authenticated session.

How it works

It uses a Chrome extension to bridge the gap between an external controller (CLI or MCP client) and the browser. The system consists of a WebSocket relay server running on localhost that communicates with the extension via the chrome.debugger API. This allows the agent to use the full Playwright API to interact with the browser without requiring the browser to be launched with special remote-debugging flags that are often blocked on default user profiles.

Who it’s for

  • AI Agent Developers: Those building agents that need to perform tasks in a user's authenticated browser session.
  • Automation Engineers: Users who want to programmatic control over their own browser with existing state.
  • Remote Operators: People who want to share a specific browser tab with a remote agent or person for authenticated work.

Highlights

  • Existing Session Access: Uses your real Chrome profile, including logins and extensions.
  • Full Playwright API: Provides a complete automation surface rather than a limited set of predefined tools.
  • MCP Support: Integrates with the Model Context Protocol (MCP) for easy agent adoption.
  • Stateful Execution: Maintains state across multiple CLI calls via a state object.
  • Visual Labels: Generates accessibility snapshots with Vimium-style labels to help agents identify and click elements.
  • Remote Control: Allows sharing a single tab with a remote agent via a secure tunnel.
  • CDP Access: Provides first-class access to the Chrome DevTools Protocol (CDP) for advanced debugging and live code editing.

Related

  • Project
    anomalyco/browser-control

    Browser Control is an open‑source Node.js driver that lets AI agents or CLI users run Playwright code inside their own Chromium browser (with all logged‑in sessions and extensions). It works via a local relay and a browser extension, supports persistent sessions, inspection helpers, human hand‑offs, video/ HAR recording, and a TypeScript Effect client for secret‑safe API calls. Intended for trusted local use only.

  • Project
    AgentDeskAI/browser-tools-mcp

    An MCP server and Chrome extension that gives AI agents access to a user's active browser session for real-time debugging, network inspection, and Lighthouse audits.

  • Project
    SawyerHood/dev-browser

    A browser automation tool that allows AI coding agents to maintain persistent browser sessions and use compact accessibility trees for efficient web interaction.

  • Project
    BetterWright/betterwright

    A persistent browser runtime for AI agents that allows them to maintain state and interact with the web using Playwright, featuring human handoff and MCP support.

  • Project
    sleepinginsummer/agent-browser-cli

    A CLI tool that connects AI Agents to a real Chrome browser session, allowing them to reuse login states and perform precise page control and perception.