Browser automation
LibrAgent agents can open web pages, read content, and capture screenshots.
Built-in agent browser (default)
By default, LibrAgent opens a dedicated agent browser.
- It runs in a separate process, so a heavy or broken page does not freeze the main app.
- Logins you make inside that browser stay on this device until you clear them under Settings → System → Agent browser → Clear agent browser data.
- This profile is not your everyday Chrome.
Use it for public search, docs, and sandbox work.
Your Chrome (only when you need it)
Install the Chrome extension when the agent must use sites where you are already signed in — carts, mail, intranet, personal dashboards.
| Built-in agent browser | Your Chrome | |
|---|---|---|
| Logins | LibrAgent-only profile | Your normal Chrome cookies |
| Best for | Public research | Personal or work sites with your session |
| Setup | None | One Chrome extension install |
Install from the Chrome Web Store
- Start the LibrAgent desktop app.
- In Chrome, install LibrAgent Browser Bridge.
- Open the toolbar icon and wait until the status shows Connected.
If you see “LibrAgent not running”, start the app and wait a few seconds. - Ask in chat, for example:
Check my shopping cart in the Chrome window I’m using.
Settings → System → Your Chrome also has an Install from Chrome Web Store button.
Load from a local folder (developers)
Copy the path from Settings, then in Chrome open chrome://extensions, turn on Developer mode, and choose Load unpacked.
NOTE
- After install, the extension reconnects when you restart LibrAgent.
- Each chat keeps one active browser. Switching modes closes the previous link.
- The extension only links LibrAgent and Chrome on this computer. It does not upload your browsing data to a LibrAgent cloud by itself.
What you can do
| Capability | Built-in | Your Chrome |
|---|---|---|
| Open and move between pages | ✅ | ✅ |
| Read page content | ✅ | ✅ |
| Click, type, scroll | ✅ | ✅ (works on many strict-security sites) |
| Screenshot | ✅ (visible area or full page) | ✅ (visible area only) |
| Console logs | ✅ | ❌ (not yet) |
| Use your Chrome logins | ❌ | ✅ |
Typical “Your Chrome” tasks: check logged-in shopping, mail, or GitHub; open intranet dashboards; capture the visible tab; click and type in forms.
Example uses
- Live docs and news — Open public pages and summarize them.
- Local UI checks — Open
http://localhost:3000and capture the screen. - Scheduled briefings — Pair with Scheduled Tasks to visit the same sites each day.
Note
- The first launch of the built-in browser may take a few seconds while the runtime starts.