↓

Write a Zombie Couple Video Script

Oct 29, 2026

Scope: The creative examples below describe general video-making ideas, sometimes using other tools or editing. The current Zombie Couple Video homepage offers one fixed reunion story from two adult photos, with swappable roles and a fixed audio reference. It does not offer custom prompts, a scene catalogue, model selection or added captions. AI likeness, clothing, scenery and audio timing may vary.

You don't need a script for ten seconds. You need a beat.

The word "script" makes people write dialogue, which is the one thing this format can't do. What it can do is one small action with a readable reaction, and that's a different craft — closer to writing a sight gag than to writing a scene.

One beat, two parts

A beat is: something happens, and someone notices.

That's it. Two parts. The first part is the action; the second is the response. Everything that works in ten seconds has both.

Examples of the structure:

  • A hand reaches out → the other person looks at it
  • A door opens → both faces turn
  • A mug is offered → a pause → it's taken

Notice that each of these is describable in a single sentence and contains a change. A scene with no change isn't a beat; it's a mood, and a mood doesn't survive ten seconds of viewer attention.

The reaction is the clip

This is the part people get wrong. They design an action and forget that the action alone isn't interesting — the response is what the viewer reads.

Two consequences:

Frame for the reaction. If someone turns toward a sound, the frame should see their face when it happens, not their back.

Give the reaction time. Half the ten seconds for the action, half for the response. An action that resolves in two seconds leaves eight seconds of nothing.

Three worked examples

The mug

  • Beat: one person offers a mug, the other takes it.
  • Prompt: "One person hands a mug to the other, both faces visible, warm lamp light from the left."
  • Why it works: small movement, two faces always in frame, a prop that gives the model something concrete to hold.

The stand-off

  • Beat: two people face each other, neither moves, then one tilts their head.
  • Prompt: "Two people face each other at a short distance, still, both watching. Hard light from one side."
  • Why it works: minimal motion, maximum face time. The tilt is the whole payoff and it costs nothing.

The door

  • Beat: both look at a closed door, then turn to each other.
  • Prompt: "Two people stand at a closed door, then turn toward each other. Cold light from behind."
  • Why it works: the turn is a single readable action, and "cold light from behind" gives a rim that separates them from the background.

Dialogue doesn't work

Write it down and it becomes clear fast: any line of dialogue requires lip-sync, and lip-sync in a ten-second AI clip is where results fall apart.

The workaround is caption text, added afterwards in any editor. One line, at the end or the start. That's a caption, not a script, and it's the right tool.

If the joke requires a line, the line goes in the caption. If it requires a delivery, the joke needs a different format.

What to cut

Most first drafts of a ten-second idea contain three beats. Cut two.

Signs you have too much:

  • Your prompt has the word "then" more than once
  • You've described what happens before the clip and after it
  • You can't say what the clip is about in one sentence

The version that survives is usually the smallest one: two people, one action, one response.

The test before you spend a render

Read your prompt out loud and ask: if I only saw the last three seconds, would I understand what happened?

If yes, the beat is legible and the clip will work. If no, the payoff is buried and you'll get something that reads as aimless motion — which is what most disappointing AI clips actually are.

Try it at the generator; price at pricing.

Frequently asked questions

Do I need to write a script?

No. Write one sentence describing an action and one describing the response.

Can the characters speak?

No — lip-sync is unreliable at this length. Put any line in a caption instead.

How many things should happen?

One. Two at most if they're causally linked, like a reach and a take.

What if my idea needs more than ten seconds?

Then it's a different format. Cut it down to the strongest moment, or make two clips and accept the cut.

Should the reaction be at the start or the end?

End. The reaction is the payoff, and payoffs go last.

Zombie Couple Video