Watch and Control Agents in the Browser

Agent Browser is a real Chrome that runs on your TermSquad computer. Agents use it to open and test the apps they build while you watch the same page in the Web Terminal. It is included on every plan.

Ask an agent to use the Browser

Ask in your own words, for example: “Test the sign-up flow in the Browser and fix anything broken.” Compatible agents get Browser tools automatically. When an agent starts browsing, the Browser panel opens beside the agent terminal. The Browser button sits next to Files in the toolbar.

Because Chrome runs inside your computer, agents can open http://localhost:3000 or any other dev server running there. You do not need a public preview link.

Agents can navigate, click, type, fill forms, upload files, take screenshots, read console and network logs, check CSS, and run Lighthouse audits and performance traces. The agent’s cursor appears on the page.

Take control

While an agent is in control, the panel shows Live and “{Agent} is controlling the browser”. Select Take control to use the page yourself. The status changes to You control.

While you hold control, agents can still look at the page, but clicks, typing, navigation and scripts are blocked. Select Return control to {Agent} when you are done. Control also returns automatically after two minutes of inactivity, when the app is in the background or disconnected for a short time, or when you close the panel. Closing the panel never stops the agent’s Chrome.

On mobile

On a phone, the Browser opens almost full screen. Use ← Terminal to go back, and Type in browser to bring up the keyboard while you hold control.

Limits and privacy

Each agent gets its own Chrome with a private profile that persists between sessions. How many browsers can run at once depends on your computer’s memory. Page images and keystrokes travel only between your browser tab and your computer; TermSquad does not store them.