Window Peeper prompt: a complete 10-second scene

Copy a full image-to-video prompt, choose a usable reference photo and give the awkward pause enough time. No account or credits are needed to use this guide.

Published

01

Pick the face before the scene

One recognisable subject does more for the joke than a complicated background.

BRING THIS

  • One clear subject. A consenting adult facing the camera, with even light and an unobstructed face.
  • A readable outfit. A waist-up reference gives clothing cues; an unusual outfit can become part of the joke.
  • A recognisable pet. Show the face, coat colour and markings clearly, with room around the ears.
  • A simple crop. Leave enough of the upper body for the seated pose.

LEAVE THIS OUT

  • Group photos. Multiple faces make the intended subject ambiguous.
  • Hidden detail. Heavy blur, dark glasses, face coverings or hands over the face.
  • Someone else's identity. Strangers, public figures or people who have not agreed.
  • A promise of exact likeness. Clothing, fur patterns, hands and facial detail need review in any generated result.

Selection principles informed by Viberot's photo guide. Selecting a photo previews it locally. Only choosing Create uploads it for processing. Likeness, clothing and markings can vary in an AI result.

02

A ten-second prompt, with room for the stare

Our own writing of the scene. Copy it, adapt it to your provider, then review the result.

  1. 0–2.5 seconds

    Approach the building. Travel toward the main warm window in one continuous move.

  2. 2.5–4 seconds

    The subject takes a bite, occupied with the burger rather than the camera.

  3. 4–7.5 seconds

    Recognition, then an uncomfortable pause. Hold a flat, unamused stare.

  4. 7.5–8.5 seconds

    Set the food down and begin to stand. Let the stare continue.

  5. 8.5–10 seconds

    Draw the curtain all the way across. End on the closed, softly lit fabric.

Read & copy the complete original prompt

English · one reference · audio direction included

Make a ten-second portrait-format scene using the attached reference photo as the identity guide. The joke is a private late-night snack interrupted by an unwanted audience. It should feel awkward and faintly eerie, with a calm comic payoff.

Subject: use a single consenting adult from the reference. Keep the face, hair and visible clothing consistent across the shot. If the reference is a pet, retain its coat pattern, colouring and recognisable face; stage a playful upright pose. Do not introduce a second character. Treat identity and hand detail as review points, not as guaranteed results.

Setting: view an apartment facade from outside after dark. Make one warm dining-room window the clear destination of the shot; keep the surrounding building subdued. Inside, the subject sits close enough to the glass for the expression to read. Put a large burger, an unmarked takeaway bag and a plain food carton on a small table. A long curtain must be within the subject's reach from the start.

Camera and timing: maintain a single uninterrupted exterior viewpoint. From 0 to 2.5 seconds, move smoothly toward the warm window instead of jumping to a close crop. From 2.5 to 4 seconds, show the subject absorbed in taking a bite; eye contact has not begun. From 4 to 7.5 seconds, let the subject notice the viewer and become still, keeping a dry, unimpressed gaze while finishing the bite. Give this pause time to land. From 7.5 to 8.5 seconds, place the burger down and rise without breaking the stare. Only from 8.5 to 10 seconds, reach for the curtain and draw it completely shut. Hold the last glimpse of warm light behind the fabric.

Performance and image: restrained movements, ordinary clothing from the reference, believable glass reflections and a cool night exterior against the warmer room. Keep the person recognisable and the food easy to read. Avoid grinning, greeting, waving, startled jumps, extra people, abrupt transitions and changing faces. Keep hands anatomically plausible. Do not put lettering, watermarks or brand marks on the food packaging, building or curtain. Add any desired title or brand card later in an editor.

Sound direction, if the chosen model supports audio: soft outdoor ambience, a little packaging rustle and the final movement of fabric. Leave out dialogue, narration and music. If the model does not produce audio, use this as a separate sound-editing brief. Output settings and identity fidelity depend on the provider; review the result before posting.

Written for this site. The timing is a creative brief, not a tested provider preset. A model may miss a beat, alter likeness or omit sound. Our generator uses its own fixed scene instructions; this editable prompt is a creative reference.

04

Give the awkward pause a caption

Start with the joke, then use only the tags that fit your post.

My midnight snack now has an audience.

#windowpeeper #windowpeeperai #burgermeme

The curtain is my official statement.

#windowpeeperfilter #curtainpull #aivideo

The cat would like a little privacy with that burger.

#windowpeeper #catmeme #petvideo

A spooky window. A very ordinary dinner.

#windowpeeper #halloween #windowpeepertrend

Original caption ideas. Hashtags describe the scene; they do not promise reach. Add #capcut only if you actually used CapCut, and choose music you have permission to use.

Use the prompt in the right workflow

This editable brief is free to copy into a compatible image-to-video service. It is our own wording of the scene, not a tested preset or a copy of another site's full prompt. Provider controls, available audio and charges differ.

The generator on this site uses fixed scene instructions and accepts a permitted photo, quality and aspect ratio. Pasting this guide into an invisible prompt field is not required. If a take goes wrong, first use the six-problem review checklist. Read the trend sources for the background, rather than assuming a caption or hashtag guarantees reach.