Browser

Let your Agent perform ad hoc work in its live Browser.

Browser lets an Agent interact with the live Browser for work that is not represented by an exact
registered helper. It is useful for exploratory or changing workflows where the Agent must inspect
the current page and decide what to do next.

Before enabling Browser

You need:

• a provisioned Computer Agent
• the website signed in or otherwise ready for use
• a bounded description of the Browser work
• clear instructions describing when the Agent may use it

An Agent uses its own deployed Computer automatically. Select another Computer Agent only when the
Browser work must run on that other Agent's Computer, then add Browser in the Functions section.

Write clear instructions

Explain the intended outcome, the information the Agent may use, when Browser work is allowed, and
how it should handle a failed or incomplete result.

Test with non-destructive data before publishing. Confirm the expected result in the destination system rather than relying only on the Agent's response.

Use Computer for every registered reusable helper, including helpers that use
the Browser, and for broader multi-step work. Browser is not the function used to expose registered
helpers.

See Reusable helpers and browser automation for the complete creation and verification workflow.

Configuring delegation programmatically? See Computer Agent configuration and Agent functions.