Control Your Existing Chrome Tabs with the OpenClaw Extension

I often find it frustrating when a browser agent starts a completely fresh, empty profile. I have to log in to every service again, bypass two-factor authentication, and set up my workspace from scratch just so the agent can see what I’m seeing. It feels like double the work for half the result.

I prefer letting the agent jump right into the tabs I already have open. The OpenClaw Chrome extension makes this possible by acting as a relay. Instead of managing a separate browser instance, it attaches to your current Chrome window via the debugger API.

What You’ll Need

• An active OpenClaw installation.
• Google Chrome.
• The OpenClaw Gateway running (locally or on a remote node).

Quick Start

You can get the extension running in about five minutes. Here is the path to get it set up.

1. Install and locate the files

The extension files are bundled with the OpenClaw npm package. You need to move them to a stable location on your machine first.

openclaw browser extension install

After that, get the path to the directory where the files were installed:

openclaw browser extension path

2. Load the extension in Chrome

Open your Chrome browser and go to chrome://extensions.

• Turn on Developer mode using the toggle in the top right.
• Click Load unpacked.
• Select the folder path you copied in the previous step.
• Pin the extension to your toolbar so you can see it.

3. Attach to a tab

Open the tab you want the agent to control. Click the OpenClaw icon in your toolbar. You will see the badge change to ON. This means OpenClaw is now attached to that specific tab.

4. Run a command

OpenClaw includes a built-in profile named chrome that points to this extension relay. You can test it from your terminal:

openclaw browser --browser-profile chrome tabs

If you are using an agent tool, call the browser tool with profile="chrome".

Attach/Detach Logic

The extension does not automatically grab every tab you look at. It only controls the tabs where you have explicitly clicked the toolbar button. If you want to switch the agent’s focus to a different tab, just navigate to that tab and click the extension icon again.

Security and Sandboxing

This is a powerful tool because it uses the chrome.debugger API. When the badge is ON, the model can click, type, and read any content in that tab, including your logged-in sessions.

If your agent is running in a sandbox, the browser tool might be restricted to a “sandbox browser” by default. To let a sandboxed agent use your host Chrome via the extension, you need to update your config:

{
  agents: {
    defaults: {
      sandbox: {
        browser: {
          allowHostControl: true,
        },
      },
    },
  },
}

I recommend using a dedicated Chrome profile for this rather than your primary personal profile to keep your data safe.

Troubleshooting

If things aren’t working, check the extension badge icon:

• ON: Everything is working; OpenClaw can drive this tab.
• …: The extension is trying to connect to the local relay.
• !: The relay is unreachable.

The ! error usually means the browser relay server isn’t running. Ensure your Gateway is running locally. If your Gateway is on a remote machine, make sure you have a node host running on your local machine to bridge the connection. You can also open the extension Options page to see a more detailed connection status.

If you update OpenClaw, remember to refresh the extension files:

1. Run openclaw browser extension install again.
2. Go to chrome://extensions and click Reload on the OpenClaw extension.

Still having trouble getting the relay to connect? Ask the AI Setup Assistant.

What’s Next

• Browser Tool Overview
• Gateway Security Guide
• Setting up Tailscale for Remote Nodes