Dev Chrome

You run a MCP server that keeps a Chromium instance attached via Puppeteer, enabling tooling to inspect dev builds, capture screenshots, and drive the UI with permission-gated capabilities. This server lets you monitor and control a Chromium-based dev environment while staying in charge of what actions are allowed.

How to use

You will add a local MCP server entry that launches the Chromium-monitoring process. The server exposes tools for inspection, permission management, and UI automation, and you will grant permissions to perform specific actions like navigation, status checks, and DOM evaluation.

How to install

Prerequisites you need before installing:

• Node.js is required (LTS version recommended).
• NPM is included with Node.js.

Install the MCP server package in your project:

npm install @2sslgetlool/dev-chrome-monitor

Configure the MCP server

Add a new MCP server entry to your MCP configuration to start the Chrome-monitoring service. The example below shows how to point to the published script and set the necessary environment values.

{
  "mcpServers": {
    "dev_chrome_monitor": {
      "command": "node",
      "args": [
        "node_modules/@2sslgetlool/dev-chrome-monitor/dist/devChromeServer.mjs"
      ],
      "env": {
        "MCP_DEV_URL": "http://localhost:5173/",
        "MCP_CHROME_PATH": "C:\\\\Program Files\\\\Google\\\\Chrome\\\\Application\\\\chrome.exe",
        "MCP_NAV_TIMEOUT": "45000",
        "MCP_WAIT_TIMEOUT": "45000"
      }
    }
  }
}

Start the MCP server and verify

Reload or restart your MCP client to pick up the new server configuration. You can then inspect and manage permissions, view status, capture screenshots, and run evaluations once the appropriate permissions are granted.

Usage notes and permissions

Two primary actions govern what you can do with the Chrome-monitor MCP server: you view current permissions and you grant or revoke them. Use these tools to control access to navigation, status checks, screenshots, evaluations, and DOM automation. When a tool requires a specific permission, you must grant that permission before the tool will run.

Environment variables and defaults

The server relies on a few environment variables to set up the monitoring context. The defaults shown are typical, but you can customize them to fit your environment.

MCP_DEV_URL=http://localhost:5173/
MCP_CHROME_PATH=C:\\Program Files\\Google\\Chrome\\Application\\chrome.exe
MCP_NAV_TIMEOUT=45000
MCP_WAIT_TIMEOUT=45000

Development

If you are actively developing the MCP server itself, install dependencies and build the project to keep the dist bundle in sync with the source.

npm install
npm run build

Security considerations

Only grant permissions you truly need and review the list of capabilities before enabling actions that navigate, read, or modify the UI. Use the permission controls to restrict access to sensitive operations and ensure the dev server is exposed only to trusted tooling.

Troubleshooting

If the server fails to start, verify that Node.js is installed, the path to Chrome is correct, and the dev server URL is reachable. Check the MCP client for permission prompts and ensure you have granted the required capabilities before attempting actions like navigation or DOM evaluation.

Available tools

dev_list_permissions

Inspect which capabilities are enabled for the Chrome-monitor MCP server.

dev_grant_permissions

Grant specific permissions (e.g., status, navigate) to allow tools to run.

dev_revoke_permissions

Revoke previously granted permissions to restrict access.

dev_status

Check the current status of the MCP server and its connected Chromium instance.

dev_screenshot

Capture a screenshot of the current browser UI being monitored.

dev_eval

Evaluate or inspect the DOM or page state within the Chromium instance.