Chrome FAQs

Question 1

What is the Chrome MCP Controller?

Accepted Answer

The Chrome MCP Controller is an MCP (Model Context Protocol) server paired with a Chrome browser extension that allows AI agents to programmatically control Chrome tabs, automate web interactions, and extract information.

Question 2

How does it enable AI agents to control Chrome?

Accepted Answer

It works via a dedicated WebSocket server (port 9223) that communicates with both the AI agent (e.g., godex) and the Chrome extension. Commands from the AI agent are pushed to the extension, which then executes actions within Chrome tabs.

Question 3

How can I integrate the Chrome MCP Controller with my AI agent?

Accepted Answer

It integrates with AI agents supporting the Model Context Protocol (MCP), such as godex. You configure the `chrome-mcp` command as an MCP server in your agent's provider settings (e.g., `~/.godex/providers.yaml`), specifying the command path and transport type.

Question 4

What specific actions can AI agents perform in Chrome using this tool?

Accepted Answer

AI agents can list connected tabs, get tab information, extract structured DOM overviews, retrieve readable page content, execute custom JavaScript, find specific DOM elements, get element details, wait for elements, and capture full-page screenshots.

Question 5

Is the Chrome MCP Controller secure to use?

Accepted Answer

Yes, it is designed for secure local operation. A critical security warning is to 'Ensure port 9223 is blocked from external traffic' as it runs a local WebSocket server. This ensures that only your local AI agents can interact with it.

Chrome

Chrome

主な機能

ユースケース

主な機能

ユースケース