Your AI Agent Can Now See Your Real Browser (And This Post Was Written With It)

Source: Dev.to

Overview

Real Browser MCP is an MCP server plus a Chrome extension that gives your AI agent control of your actual browser. It works with the same tabs, cookies, and logins you already have—no headless browser, no replaying authentication for quick checks like “did the button work?”.

Traditional Playwright MCP and Chrome DevTools MCP launch a fresh browser with a new profile, which is fine for isolated end‑to‑end tests but cumbersome when you want an agent to interact with a page you already have open. Real Browser MCP solves this by running the extension inside the Chrome instance you’re using and communicating with the MCP server over a local WebSocket.

How It Works

• MCP server: Runs locally and can be started with a single command.
• Chrome extension: Installed in your existing Chrome browser; it connects to the MCP server.
• The agent can then:
  • Snapshot the page
  • Click, type, and scroll
  • Take screenshots

All actions are performed in the same browser session you’re already using.

Installation

MCP Server

npx -y real-browser-mcp

Chrome Extension

Install the extension from the Chrome Web Store:

Real Browser MCP – Chrome Web Store

After installation, a green dot in the extension popup indicates a successful connection.

Usage Example

1. Start the MCP server with the command above.
2. Ensure the Chrome extension shows the green dot (connected).
3. Your AI agent can now see and interact with the pages you have open.

Example: The author wrote this post using Real Browser MCP. The agent loaded the draft, navigated to dev.to/new, filled out the form, and published the article—all within the same logged‑in browser.

Open Source

Real Browser MCP is open source under the MIT license.

• GitHub repository:
• Chrome Web Store: