A prompt for a zombie AI video is just a plain description of the scene you want, written short enough that you can hold it in your head while you pick your photos. On this site the five effects all run on the same image-to-video pipeline, so the prompt is less about a text box and more about knowing what you are aiming for before you upload. Below are fifteen scene ideas you can copy, grouped by effect, plus a short note on why the photos still matter more than the words.
How prompts work here
It helps to be clear about the division of labor. The model on this site is an image-to-video model: it animates your photos into a short clip. The effect you pick decides the story shape, and your photos decide who is in that story. A prompt is the plan you write down before you start, so that you know which photos you need and what the clip is supposed to feel like.
That means the most useful prompts are concrete. "Zombie" alone tells you nothing about which photo to pick. "A survivor lowers a weapon and hugs a turned partner" tells you immediately that you need two faces. The prompts below are written that way, as short briefs you can keep beside you while you work.
Zombie love story prompts
The love story effect takes two photos and runs the four-beat format: the standoff, the lowered weapon, the hug, and the hard cut into a warm memory. These prompts all fit that shape.
- A survivor aims a weapon at a turned partner on a front porch, then lowers it and pulls them into a hug before the shot cuts to a summer memory.
- A couple in wedding clothes, one of them turned, the other still holding a knife, then a cut to the same pose on a bright wedding day.
- Two friends in a car at night, one turned in the passenger seat, a hand on the shoulder, then a cut to a road trip memory in daylight.
- A parent facing an infected adult child in a hallway, the weapon dropping, then a cut to a birthday party in the same room.
- A soldier and a turned partner in a ruined living room, the rifle lowered, then a cut to the same room before the outbreak, curtains open.
Zombie pet prompts
For this one you use the zombie pet effect and a clear photo of an animal. The prompt here is mostly about mood and setting.
- A dog on a porch at dusk, turned and staring at the camera, mist rolling across the boards.
- A cat on a kitchen windowsill, eyes milky, rain on the glass behind it.
- A small dog in a child's bedroom, turned but still curled up where it always slept.
- A large dog standing in a wheat field at golden hour, one eye catching the light, turned.
- A cat in a doorway, backlit, silhouetted, the room behind it dark and empty.
Halloween makeup prompts
The AI Halloween makeup effect uses a photo of a face. These prompts describe the look rather than the plot.
- A cracked porcelain look, white base with fine black cracks across the cheeks and a run of red at the mouth.
- A classic ghoul, deep-set eyes, grey skin, a torn lip and dirt along the jawline.
- A glamour horror look, glossy red lips, dark eyes, and a single line of stitching across the forehead.
- An undead bride, chalk skin, dark veins climbing the neck, and a veil casting a shadow over the eyes.
- A burnt-forest creature, ash-grey skin, glowing amber eyes, and leaf ash across one cheek.
Mini me and giant prompts
These two use the other effects on the site. The AI mini me effect puts a smaller version of a person into the frame, and the make-things-giant effect scales a subject up against where they are standing. Two briefs for each:
- Mini me: a person standing in a kitchen while a tiny version of themselves sits on the counter, both looking at the camera.
- Mini me: someone on a sofa, a thumb-sized copy of them standing on the armrest, the room otherwise ordinary.
- Giant: a person walking down a normal street, scaled up until their shoulders reach the shop signs.
- Giant: someone standing in a back garden, huge against the fence, a normal-sized chair at their feet.
How to turn a prompt into a clip
The workflow is short and the order matters. Write the prompt first, then choose the effect, then pick the photos that match. If your prompt is "a soldier lowers a rifle and hugs a turned partner," you already know you need two photos and that the love story effect is the right one. If your prompt is about a single turned pet, you need one animal photo and the zombie pet effect. Doing it in that order stops the most common frustration, which is uploading a photo and then trying to force a scene onto it.
Keep the prompts short. A clip is five seconds, so a brief that describes a whole story with a beginning, a middle and an end will not fit. Notice that every prompt above describes one moment, or one moment plus the cut. That is the right scale for this format.
Why the photos still matter more
However good a prompt sounds, the clip is only as good as the faces you feed it. A brief like "a survivor lowers a weapon and hugs a turned partner" is a plan, not a magic spell; the model still needs two clear photos to build the two characters. So the practical workflow is to write the prompt first, then pick photos that match it, not the other way around. If the brief calls for a close hug, for instance, you want two photos where both faces are clearly visible, because the hug is where the likeness matters most.
It also helps to keep one variable fixed at a time. If a clip does not land, change one thing, either a photo or the scene idea, and run it again. Changing both at once means you will not know which one helped.
Try these prompts on this site
If you want to see how one of these briefs actually looks in motion, you can run it on this site with your own photos. Pick a prompt from the list above, choose the matching effect, and upload the photos it calls for. The preview is free and watermarked, so the worst case is that you learn which of your photos works and which does not. Start with the zombie video generator if you want the love story, or go straight to AI mini me and make-it-giant for the two scale effects. When you are ready to remove the watermark, the pricing page lists the credit tiers.