Skip to content

Screenshots

Fresh

The browser subagent can take screenshots of pages and individual elements. Screenshots are surfaced as artifacts that you can review and comment on.

How to Trigger

  • Automatic -- The browser subagent takes screenshots during its operations
  • Manual prompt -- Ask the agent to "take a screenshot of the current page" or similar
  • Tool invocation -- Screenshots are available as a tool the subagent uses during browser operations

Features

  • Saved as image artifacts in the conversation
  • Support commenting for providing feedback to the agent
  • Can capture full pages or specific elements

TIP

Use screenshots to verify visual changes. Comment directly on a screenshot to tell the agent what needs to change -- "the button should be blue, not green" is much clearer than describing the issue in text alone.

SOP Documentation Site