A Chrome extension (Manifest V3) that connects your local Chrome profile to OpenAgent Browser Use, letting AI agents operate tabs you already own — complete with your logins, cookies, and extensions.
The extension runs a persistent WebSocket connection to your local OpenAgent backend. When Browser Use issues a command, it travels through that connection to the extension, which executes it in the target tab:
OpenAgent backend ──WebSocket──► background.js ──chrome.tabs──► content.js ──DOM──► page
Supported commands: tabs, state, open, snapshot, click, type, press, switchTab, closeTab, playMedia, mediaState.
Search for "OpenAgent Browser Bridge" in the Chrome Web Store and click Add to Chrome.
- Clone this repository:
git clone https://github.com/the-open-agent/openagent-chrome.git
- Open Chrome and go to
chrome://extensions - Enable Developer mode (top-right toggle)
- Click Load unpacked and select the cloned folder
- Make sure your OpenAgent backend is running (default:
http://127.0.0.1:14000) - Click the extension icon in the Chrome toolbar
- Enter the OpenAgent URL (leave as default if running locally)
- Optionally enter an auth token if your backend requires one
- Click Connect
The status indicator turns green when the bridge is live. The extension automatically reconnects if the connection drops.
| Permission | Why it's needed |
|---|---|
tabs |
List and switch between tabs |
activeTab |
Access the currently active tab |
scripting |
Inject the content script to interact with page DOM |
storage |
Persist server URL and connection preferences |
webNavigation |
Enumerate frames within a tab for iframe support |
alarms |
Keep the service worker alive for a persistent connection |
<all_urls> |
Operate on tabs opened to any URL |
Sensitive Chrome-internal pages (chrome://, chrome-extension://, devtools://, etc.) are always blocked from AI control.
See PRIVACY.md.
Pull requests are welcome. Please open an issue first for significant changes.
Apache 2.0 — see LICENSE.