For AI agents: the complete documentation index is available at /next/llms.txt, the full documentation bundle is available at /next/llms-full.txt, and this page is available as Markdown at /next/ai/agent-lynx/screenshots.md.
AI logo
AI
  • English
  • Screenshots

    Connect an app and select the target session before capturing images.

    Annotated and Fullscreen Screenshots

    agent-lynx screenshot --annotate -o ./page.jpeg

    This captures the page and refreshes the snapshot. Image labels such as [3] correspond to refs such as @e3, so an agent can identify a visible target and interact with it. Add --json to receive the screenshot path, dimensions, and the same fresh snapshot as structured output.

    The diagram below uses a real capture of the Tailwind CSS example running in Lynx Explorer. Hover over, select, or use the keyboard to focus a connected target to highlight both its annotation and its snapshot ref.

    To capture the full screen, use agent-lynx screenshot --fullscreen -o ./screen.jpeg. Use --annotate and --fullscreen separately.

    Direct Screenshots and Scrollable Content

    agent-lynx take-screenshot --fullscreen -o ./screen.jpeg
    agent-lynx take-content-screenshot --selector '#feed' --format png --scale 1 -o ./feed.png

    take-content-screenshot captures the first matching scroll container, including content outside the viewport. Use a scroll-view or a compatible list; the runtime must support the content-screenshot UI method on that element. Use screenshot --annotate when the image needs element-ref labels.

    Except as otherwise noted, this work is licensed under a Creative Commons Attribution 4.0 International License, and code samples are licensed under the Apache License 2.0.