Let AI agents see and control this browser: DOM snapshots, screenshots, clicks, typing, navigation, console and network.
Users
8
+3 in 30 days
Rating
0.0 / 5
unchanged in 30 days
Ratings
0
0 in 30 days
Store rank
#212K
#178K among extensions
Version
0.1.23
updated 3 Sep 2026
Users
monthly, 2 months
Latest, 16 Oct 2026:8 usersMove the mouse over the chart for any month
Ratings count
monthly, 2 months
Latest, 16 Oct 2026:0 ratingsMove the mouse over the chart for any month
The Chrome Web Store rounds user counts down to the leading digit, so the user line moves in steps. The ratings count is exact and is the better signal of real adoption.
About Deeporax Browser MCP
Deeporax Browser MCP connects your browser to AI assistants that speak the
Model Context Protocol, so an assistant can read the page you are on and act
on it the way you would.
It drives the browser you already use, with your profile and the sessions you
are already signed in to, instead of launching a separate automated browser
that knows none of that.
What an assistant can do once connected:
- Read the page as a structured outline of its headings, links, buttons and
form fields, including content inside web components that a normal script
cannot reach
- Capture the visible area or the whole page as an image
- Click, type, select options, drag and press keys. Input is delivered through
the browser's own input pipeline, so pages built on component libraries
respond to it exactly as they respond to a person
- Navigate, manage tabs, and move back and forward through history
- Read console output and network activity, which is useful when debugging a
site you are building
- Fill forms field by field, with every field read back afterwards so a value
that did not take is reported rather than assumed
While a tool is acting, an overlay on the page shows what is happening and
gives you a stop button.
How it works
The extension does nothing on its own. It connects to a small server you run
on your own machine, and only that server can send it instructions. Nothing
happens unless you start the server and ask your assistant to do something.
Page content and screenshots are sent to the assistant you connect, in the
same way anything you paste into that assistant is. Choose what you point it
at accordingly, and use the stop button whenever you want it to halt.
Chrome will show a banner saying the extension has started debugging the
browser. That is expected: it is how the extension delivers real input and
takes screenshots, and it is Chrome telling you the truth about what is
happening. Dismissing it stops the extension from acting.
The project is open source. Setup instructions are in the repository linked
below.