The October 2026 trend · two photos · two prompts
Zombie Couple Video: the two prompts behind the trend
One photo of him, one photo of her. Part 1 turns her zombie and ends in a mid-air embrace; part 2 replays that exact pose as a golden-hour memory. Copy the verbatim prompts, paste them into Google Flow, and stitch the two clips in CapCut.
How to make a zombie couple video
Two photos, two prompts, one stitch — no editing skills needed.
Prepare two photos
One clear, front-facing photo of each person. Same order in both prompts: Reference Image 1 and Reference Image 2. Pets work too — one owner, one dog.
Pick a scene and copy part 1
Choose from 8 scenes on the prompts page. Copy the horror prompt, attach both photos, generate. It ends frozen in the embrace pose.
Match the last frame with part 2
Copy the memory prompt, re-attach the SAME two photos in the SAME order, generate again. It opens on the frozen pose and melts into golden hour.
Stitch in CapCut and post
Cut the two clips at the frozen embrace so the pose lines up, add a licensed sound, and post it. The stitch is invisible when the pose matches.
The original snow-cabin prompts, verbatim
This pair started the trend on October 3, 2026. Copy them exactly — every line, including the negative list at the end that keeps the clip platform-safe.
Snow Cabin (the original)
The scene that started the trend: an abandoned snow-covered mountain cabin, a tearful standoff, a mid-air embrace — then a golden-hour memory on the same slope.
Create a 10-second photorealistic cinematic video using the 2 uploaded reference images. Reference Image 1 = MAN and Reference Image 2 = WOMAN. Preserve both faces, identities, facial features, and hair exactly from the references. Create a scene inside an abandoned snow-covered mountain cabin. The man wears a beige beanie and petrol-blue winter parka and is crying while aiming a handgun at the woman. The woman is transformed into a terrifying zombie with pale grey-blue cracked skin, milky white eyes, messy hair, a torn dirty beige dress, and bare feet. The man slowly lowers and drops the gun, closes his eyes, and spreads his arms. The zombie woman suddenly runs toward him and leaps into his arms. She wraps her arms around his neck in mid-air as both bodies fall backward together. End with them frozen in the mid-air embrace pose. Photorealistic live action, cinematic 35mm lens, realistic physics, shallow depth of field, natural film grain, cold blue winter lighting. Only these two people. No dialogue, text, subtitles, logos, blood, gore, biting, or extra people.
Use the SAME TWO uploaded reference images again. Reference Image 1 = the SAME MAN. Reference Image 2 = the SAME WOMAN. Keep both identities, faces, and facial features exactly consistent with the reference images. Start by matching the final mid-air embrace pose from Part 1. Seamlessly transform the cold cabin into a beautiful golden-hour snowy mountain slope. The SAME woman from Reference Image 2 is now healthy and beautiful, wearing a white shearling coat, black pants, and fur-trimmed boots. Continue their fall into deep snow in graceful slow motion. They land together, laugh, and embrace. Then show them running hand-in-hand through the snow toward the sunset, embracing on a hilltop. End with the SAME man gently kissing the SAME woman's forehead while the golden sunset glows behind them. Photorealistic live action, cinematic 35mm look, realistic snow physics, shallow depth of field, warm golden backlight, soft film grain, emotional dreamlike atmosphere. Only the SAME two people. No dialogue, text, subtitles, logos, extra people, blood, or gore.
8 scenes built on the same structure
The scene sentences swap; the identity rules, the frozen-pose bridge and the negative list never change.
Why your zombie video got blocked
Almost every rejected clip fails for one of four reasons — all four are fixable in the prompt.
Faces drift between parts
Fix: name each reference image, demand exact identity preservation, and reuse the SAME two photos in the SAME order in part 2.
Gore words trigger moderation
Fix: keep the negative list at the end of every prompt — no blood, no gore, no bite marks. Horror mood comes from lighting, not wounds.
Text overlays and logos
Fix: the prompts forbid on-screen text and logos. Add captions later in CapCut, never inside the generation.
Copyrighted music
Fix: generate silent, then add a licensed or trending sound in your editor or the TikTok app itself.
Everything else on this site
Each page targets one thing people search for around this trend.
Frequently asked questions
The six questions people ask most about the zombie couple video trend.
It is a two-part AI video made from just two photos. Part 1 turns one person into a zombie in a horror scene that ends with a mid-air embrace; part 2 starts from that exact pose and transforms the same place into a golden-hour memory of the couple. The trend started on October 3, 2026, when a two-prompt Google Flow tutorial went viral.
Prepare two clear photos, one of each person. Copy the two prompts for a scene from this site, upload both photos in the same order to Google Flow or any image-to-video tool, generate part 1, then generate part 2 in a new run with the same photos. Stitch the two clips in CapCut so the frozen embrace pose lines up.
Three common causes: faces drifting between the two parts (fix by keeping the same reference images in the same order), gore descriptions like blood or bite marks (the prompts here already end with a negative list that removes them), and copyrighted music or on-screen text. Swap licensed tracks in CapCut, not in the AI tool.
The trend was built on Google Flow, but the same two-prompt method works in any image-to-video model that accepts reference photos, such as Kling, Veo or Seedance. Use the same pair of prompts and the same two photos in every tool; only the stitching step stays in your editor.
Name each reference image in the prompt (Reference Image 1 = MAN, Reference Image 2 = WOMAN), require faces, identities and hair to be preserved exactly, and repeat the same two images in the same order in part 2. Ending part 1 frozen in a specific pose and starting part 2 from that pose is what hides the cut.
Yes. The Parent & Child scene replaces the zombie makeup with a tired expression and torn but clean clothes and removes all gore words, which keeps it within platform policy for family audiences. The Owner & Dog scene needs no horror elements at all and performs reliably on TikTok.