Three copy-paste prompts for the viral zombie trend: the original "can't shoot" video, the zombie love story and the zombie hug. Paste one into an image-to-video AI tool that accepts two reference photos.
They're simple starting points and haven't been tested in every tool: results vary, so expect to adjust them and try a few times.
Pick an image-to-video AI tool that takes two reference photos.
Upload one clear, front-facing photo of each person. Person 1 is the survivor; Person 2 is the one who turns. If your tool names its reference images differently, rename them in the prompt.
Choose vertical 9:16 and 10–15 seconds, paste the prompt and generate. If a face drifts, generate again; a shorter clip can help.
The prompts
Three zombie trend prompts
The original: can't shoot
The gun won't fire, a warm memory flashes by, and a hug instead of a bite.
Vertical 9:16 video, about 15 seconds, photorealistic and cinematic, in dark, cold colours.
Two people from the reference photos. Person 1 is the survivor. Person 2 has turned into a zombie: pale grey skin, tired clouded eyes, torn clothes, but still clearly recognizable. Keep both faces true to their photos.
Shot 1: On an empty street at dusk, Person 1 points a handgun with shaking hands, crying. The gun does not fire.
Shot 2: Person 2 stumbles out of the shadows toward them.
Shot 3: A short warm flashback: the two of them laughing together in golden sunlight, both fully human.
Shot 4: Back on the street, Person 1 lowers the gun, closes their eyes and opens their arms.
Shot 5: Person 2 hugs them instead of attacking. Hold on the embrace.
Music: slow, sad piano. No blood, no gore, no gunshot, no text on screen.
No gun: they recognize each other, memories come back, and the bond holds.
Vertical 9:16 video, about 15 seconds, photorealistic, a quiet romantic short film. The present looks grey and washed out; the memories look warm and golden.
Person 1 is the survivor. Person 2 has turned into a gentle zombie: pale grey skin, faded eyes, worn clothes. Person 2 moves slowly and is still clearly recognizable. Keep both faces true to their photos.
Shot 1: In an overgrown, empty park, Person 1 finds Person 2. They stare at each other; Person 1's eyes fill with tears.
Shot 2: Person 2 slowly reaches out and takes Person 1's hand, as if remembering.
Shot 3: Flashback, both fully human: singing together on a road trip, then sharing a picnic in the sun.
Shot 4: Back in the present, they sit side by side holding hands as a ray of light breaks through the clouds.
Music: soft piano. No weapons, no blood, no text on screen.
No gun: arms open, and in the hug the one who turned becomes human again.
Vertical 9:16 video, about 12 seconds, photorealistic and cinematic.
Person 1 is human. Person 2 has turned into a zombie: grey skin, clouded eyes, torn clothes, but still clearly recognizable. Keep both faces true to their photos.
Shot 1: On a foggy road, Person 2 staggers toward Person 1 with arms reaching out.
Shot 2: Person 1 doesn't run. They stand still, close their eyes and open their arms.
Shot 3: They hold each other tightly. During the hug the grey slowly fades, and Person 2 looks like themselves again.
Shot 4: The two smile at each other in warm evening light.
Music: gentle strings. No weapons, no blood, no text on screen.
Any image-to-video tool that accepts two reference photos. We haven't tested these prompts in every tool, so expect to adjust the wording and try a few times. ZombieCouple.tv itself runs on ByteDance's Seedance 2.5 and needs no prompt.
No. These are simplified starting points written for this page. The generator uses its own longer prompts, tuned for each scene and length, which is why you only upload two photos.
Yes. Copy them, change them and use them however you like.
Video models can drift away from the reference photos, especially in longer clips with many shots. Use one clear, front-facing, well-lit photo per person, avoid sunglasses and heavy filters, and keep the clip short.
Upload two photos to the ZombieCouple.tv generator and pick a scene: the story, camera and music are preset. One 12-second 720p HD video costs $9.99, paid once, and a failed render gives its credits back.
Only with their permission, and only of adults who agreed to appear.