Browser MCP App

An MCP App that drives a real Chromium browser with Playwright and renders a live, clickable view of the page inside an MCP host such as Archestra. The agent can browse for you, and you can take over in the panel — click, type, scroll, navigate.

How it works

Browser tools (browser_navigate, browser_click, browser_type, browser_press, browser_scroll, browser_back/forward/reload, browser_read_page) drive one shared Playwright page. The UI panel polls an app-only browser_screenshot tool (~1/s), maps your clicks back to viewport pixels, and forwards your keystrokes — so the model and you drive the same page.

Setup

npm install
npm run browsers   # one-time: download Chromium
npm run build      # bundle the UI into dist/mcp-app.html

Run

npm run serve         # HTTP transport on :3001  (npm run dev = watch mode)
npm run serve:stdio   # stdio transport

To use it from an MCP host (e.g. Archestra), build first and point the host's MCP config at the stdio entry:

{
  "mcpServers": {
    "browser": {
      "command": "npx",
      "args": ["-y", "tsx", "/absolute/path/to/browser-mcp-app/main.ts", "--stdio"]
    }
  }
}

Configuration

Development

Built with the create-mcp-app Agent Skill, which covers the MCP Apps SDK patterns used here (tool + UI-resource registration, app lifecycle, host context, polling). Use it when extending the server — invoke /create-mcp-app.

Notes

• One shared browser page; actions are serialized so calls don't race.

• The live view is a screenshot stream — page pixels never go to the model; only browser_read_page returns text, on request.

• It's a real browser: it can log in and submit forms. Don't point it at sites where an unintended click would be costly.