Based on the website details, Chrome MCP effectively bridges AI agents to live Chrome sessions. The local server design keeps data secure while allowing tools to navigate, click, and type. It supports existing profiles and extensions, which is ideal for QA teams. The ability to verify UI flows in a visible browser seems like a strong feature for developers who need reliable, authenticated automation without complex remote setups.
Questions about Chrome MCP
What is Chrome MCP?
Chrome MCP is a Chrome browser extension paired with a local server that allows AI agents to control a user's existing Chrome browser through the Model Context Protocol. It exposes browser tools covering live sessions, tabs, extensions, cookies, and other profile state, connecting to MCP-compatible clients via a local bridge. The project is distributed as a downloadable extension package alongside an npm package and is maintained on GitHub.
Who is Chrome MCP built for?
The tool is designed for users who want to connect automation to the Chrome profile they already use, including sessions, cookies, extensions, and tabs. It works by pairing the extension with a Chrome MCP server on the user's machine so agents can inspect pages, click controls, type into forms, and coordinate tabs without brittle scraping. This setup allows agents to work with live sessions, tabs, and practical browser tools within the real browser.
How do I get started with Chrome MCP?
To get started, users must first load the Chrome extension once by downloading the release zip, unzipping it, and loading the unpacked folder in Chrome via Developer mode. After pinning the extension, users ask their agent to install and configure Chrome MCP using a specific prompt that sets up the recommended stdio config. Finally, they open the extension popup and click Connect when needed to begin working in their live Chrome.
What specific browser tools does Chrome MCP offer?
Chrome MCP provides over 20 browser tools that allow navigation, screenshots, DOM reads, keyboard input, element actions, network-aware waits, and tab management from one browser control layer. The extension plus local server bridge ensures that sensitive browser context stays on the user's machine while agents interact through a clear local server boundary. This architecture supports practical page operations like running navigation, DOM reads, and element actions without requiring a remote browser farm.
Which MCP clients are compatible with Chrome MCP?
The system is built specifically for MCP clients such as Claude, Cursor, Codex, and other MCP-compatible agents. It exposes browser tools to these clients while keeping the working browser visible and debuggable. Users can plug into coding agents and desktop AI tools using either stdio or HTTP config, which eliminates the need for a remote browser farm.
What are the exact steps to install and configure Chrome MCP?
The installation process requires loading the extension once, asking the agent to install and configure Chrome MCP, and then clicking Connect in the extension popup. The quick start guide on GitHub provides step-by-step instructions for this two-step setup. Users can also view a live demo where an agent searches for a GPT Image, opens a real prompt, imports it into Open Prompts, and categorizes and submits it.
How does the local bridge between the extension and server work?
The extension and local server work together to turn the browser the user already uses into a controllable workspace. This allows agents to work with live sessions, tabs, and extensions directly within the real Chrome profile. The local-first control ensures that the browser state remains on the user's machine while agents interact through a clear local server boundary.
Where can I find the source code and documentation for Chrome MCP?
Chrome MCP is open source and maintained on GitHub, with documentation including a quick start guide, install instructions, and a demo of an agent-driven workflow. The project is distributed as a downloadable extension package alongside an npm package. Users can view the full source code and contribute to the development on the GitHub repository.
What is the official website of Chrome MCP?
The official website of Chrome MCP is chrome-mcp.com.
How is Chrome MCP rated on Trustburn?
On Trustburn, Chrome MCP has a rating of 4.2 out of 5 based on 5 reviews.
AI-generated from reviews on this page. Verify anything important with the company directly.
Chrome MCP presents a compelling solution for AI agents that need to interact with a real browser environment. The ability to keep sensitive context local while exposing tools like DOM reads and form submissions is a significant advantage for security-conscious teams. The local-first approach ensures that agents can operate within your existing Chrome profile, utilizing your sessions and cookies without requiring a remote browser farm. This visibility is crucial for debugging and QA workflows, as the browser remains inspectable rather than disappearing into a black box. However, the documentation suggests that while the concept is robust, the initial configuration via prompts might require some technical familiarity. It is not immediately plug-and-play for non-technical users, but for developers and QA teams, the flexibility and control offered by the local bridge seem to justify the learning curve. The integration with common MCP clients also adds to its practical value.
The installation process seems straightforward. You load the extension and let the agent handle the rest.
From the information available, Chrome MCP appears to be a solid choice for teams needing visible browser automation. The local bridge keeps context on your machine. However, the specific pricing structure is not clearly detailed on the site. It seems worth exploring if you need authenticated sessions.
Chrome MCP offers a local bridge that connects AI agents to your live Chrome session. It keeps sensitive data on your machine while allowing tools to inspect pages and submit forms. The setup involves loading an extension and using a prompt to configure the client. It seems useful for QA teams needing visible debugging and for developers who want authenticated browser context without remote farms. The information available suggests a practical solution for real-browser automation that stays under your control.
News
No news yet. This is where the business shares its own updates and press.
Photos
No photos yet. Photos shared by reviewers will show up here.
Widgets
Get the widget for your site →About Chrome MCP
Chrome MCP provides a Chrome browser extension and a local server that together allow AI agents to control a user's existing Chrome browser through the Model Context Protocol. It exposes browser tools covering live sessions, tabs, extensions, cookies, and other profile state, connecting to MCP-compatible clients via a local bridge. The project is distributed as a downloadable extension package alongside an npm package and is maintained on GitHub. Documentation includes a quick start guide, install instructions, and a demo of an agent-driven workflow.
- Website
- chrome-mcp.com