Skip to content

Can my agents use a web browser?

No. Agents cannot drive a web browser; the assistant can. Here is why unattended runs are left out of it, and what to do with the browser part of a job instead.

Last updated September 01, 2026

On this page

No. Agents cannot drive a web browser. The assistant can, and it is the only place browser work happens.

An agent asked to do something that needs a browser will not fail quietly. It says in its deliverable what needed doing in a browser, and hands that part back to you.

Why agents do not get a browser

An agent runs unattended — on a schedule, in the background, with nobody sitting in front of it. Almost every real session in a browser runs into something that needs a person:

  • A bot check that has to be cleared by hand.
  • A site asking to sign in, or a saved sign-in that has quietly gone stale.
  • A page that has changed shape since the job was first set up, so the next step is no longer where it used to be.

None of that can be handled by something with nobody watching it. Rather than half-support browser work, agents get no browser at all.

What to do instead

  1. Ask the assistant to do the browser part in chat. It opens a real browser in a panel next to the conversation, you can watch it work, and you can take the controls whenever it needs a hand.
  2. Leave the rest of the job with your agent. Everything that does not need a browser carries on as normal.

There is no setting that turns browser use on for an agent, on the agent's page or anywhere else. It is not a permission that has been left switched off — it is simply not something agents do.

An agent cannot pick up where your browser left off

This one is worth being clear about, because it sounds as though it ought to work. It does not. An agent cannot open, continue or inherit a browser session. If you leave a browser part-way through a task, the thing that carries on with it is the assistant, with you there.

Saved sign-ins stay with the assistant

Agents cannot see which sites you have saved sign-ins for, and cannot start a sign-in. That list is yours.

Nobody outside your account can drive it

The browser always runs as the account owner, and on sites where you have saved a sign-in, as you already logged in.

So somebody who reaches your assistant from outside — a caller on your phone line, or a person in an email conversation your assistant is handling — cannot make it open a browser or use your signed-in sites. Browser use is not offered to them at all.

Was this page helpful?
Still need help? Contact support