Screen Recording for an Agent Request

Show the changing behavior behind your request, with a short clip and an explanation of what matters.

Summary

Some questions depend on what happens between two screen states. A menu opens and disappears, or a dialog changes after a click. A short screen recording lets you show that sequence while explaining the problem. In supported builds, Denker’s held recording captures the display under your pointer during an active request. You can replay the saved clip and ask about the relevant changes, with the agent’s visual context depending on the chosen provider.

Key points

  • Show the transition your question depends on. A clip can preserve the action and resulting state that would be difficult to explain with one screenshot.
  • Pair the sequence with a specific question. Your narration tells the agent which interaction matters and what you want help understanding or changing.
  • Keep useful visual evidence with the request. Replay the saved clip and remove unrelated captures so the context stays focused on the task.

Explain a menu that closes unexpectedly

For example, a menu closes as you move toward an option. Start an on-screen request, capture the menu opening and disappearing, and describe what you did. Ask for help understanding that visible behavior or preparing a focused coding request.

The sequence makes the question concrete: what opened, where you moved and what happened next. Describe those actions explicitly. The agent can receive sampled images and timing context; fine motion or intermediate frames may need your own inspection of the saved clip.

This menu closes when I move toward the second option. I expected it to stay open. Help me understand the visible behavior and describe the change needed.

Capture the action and its result

Where held recording is available, hold Caps Lock during an active text or voice request to capture the display under your pointer. Release it to stop and save the clip. Each clip is limited to 20 seconds, with up to eight in one request.

Include the interaction immediately before the change and its resulting state. The saved screen video has no audio, so add your explanation through the request’s voice or text input. Drawing during capture is optional.

  1. Open an on-screen request and put the pointer on the relevant display.
  2. Hold Caps Lock while showing the menu interaction, then release it.
  3. Wait for the saved clip, replay it and remove captures you do not need.
  4. Describe the action, expected behavior and question before submitting the request.

Use the recording control available in your app

Complete the desktop setup for screen access and check that your installed build offers held recording. On Mac, Caps Lock capture also needs Input Monitoring. This shortcut expects an unremapped Caps Lock key.

What the agent can inspect depends on the provider and attachment capacity. Ask about the visible sequence you supplied and add a still image or description when a detail needs closer attention.

Share the saved clip when someone else needs it

The local recording gives you a clip to replay or share alongside a written explanation. If your dictation setup offers recording attachments, check the receiving app’s supported file route. Upload the saved file manually if that destination needs it.

Open the attachment in the destination before sharing the report. A local path identifies a file on your computer; the recipient needs the accepted attachment or an accessible shared file.