The TikTok zombie love story, three ways: one step on this site, a CapCut or Dreamina template, or a DIY build in an image-to-video model with the prompts below.
Updated October 2026.
A real sample made on this site. Yours stars your own photos.
One person per photo, face clearly visible. Photos are deleted after 7 days and never used for training.
No account needed. Pay once (plus tax where it applies), your video starts right after.
The zombie couple video is a short AI clip that spread on TikTok and Reels in autumn 2026, also called the zombie love story or zombie hug trend. It always tells the same story in about 10 to 15 seconds: one person has survived an outbreak, the person they love has turned, and the survivor raises a gun but recognises the face and can't shoot. The gun drops, the two hug, and the scene cuts to a warm memory of them before everything went wrong.
Nobody films it. People take one ordinary photo of each person and let an AI video model act out the scenes. What makes a good one is that both faces stay recognisable and the story lands: the standoff, the turned partner, the hesitation, the hug and the memory.
Pick by how much time you want to spend and how much control you want:
| Way | Time and prompts | Full story built in? |
|---|---|---|
| 1. TheZombieCouple.com | About 5 minutes, no prompts | Yes |
| 2. CapCut or Dreamina template | 10 to 30 minutes, a short prompt at most | Depends on the template |
| 3. Kling, Runway or Seedance (DIY) | An hour or more, one prompt per shot | Only if you build it |
This site exists because the DIY route takes several tools. Here the five-shot story is fixed, so you only supply the faces:
CapCut and Dreamina (ByteDance's AI tools) have zombie couple templates and an AI zombie couple page. Search the templates for zombie couple or zombie love story, add your photos and export. TikTok videos made from a template often show a CapCut · Try this template link above the caption, which takes you straight to that exact template.
The catch is that templates vary a lot. Some only turn a photo into a zombie face and stop there; some include the hug but not the memory at the end. Check the preview before you spend time on it. Free credits on these apps change often, so check the current limits in the app.
For full control, use an image-to-video model that accepts reference photos, such as Kling, Runway, Hailuo or Seedance. Give it one photo of each person as references, generate the story one shot at a time (or in one 10 to 15 second clip if the model follows multi-shot prompts well), then join the clips and add music in an editor.
These are the five shots this site uses, written as prompts you can copy and adapt. Keep the same setting in every shot, and say keep both faces exactly as in the reference photos in each one:
In an empty suburban street on Halloween night, jack-o'-lanterns glowing, fog and a full moon, the survivor (IMAGE 1) aims a handgun, alone and waiting.Close-up of the partner (IMAGE 2) as an infected: pale grey skin, milky white eyes, dark veins, but clearly the same face. No blood, no open wounds.Close-up on the survivor: recognition, tears, trembling hands; the gun slowly lowers and drops to the ground.The partner, still infected, stops a few steps away; a flicker of recognition; they run into each other's arms and hold on tight.Hard match cut in the middle of the embrace to the same two people, healthy and smiling, spinning in a sunlit wheat field. The memory fills the last 40% of the video.If you want it done tonight and don't mind paying a few dollars, use way 1. If you have time and want it free, try a template first (way 2) and fall back to the DIY build (way 3) when the template doesn't tell the whole story. For a cost breakdown of the free routes, see free and cheap zombie couple video options. For a seasonal version, see the Halloween zombie video page.
Yes, with a CapCut or Dreamina template or with free credits in an image-to-video app, if the credits last through the retries. This site is not free: one video is $6.99.
Many: Kling, Seedance (Dreamina), Hailuo and Runway all appear in the trend. This site uses Seedance 2.0 with a fixed five-shot script.
Use two separate photos, one person each. With both people in one photo, models often swap or merge the faces.
Usually the photo: a small face, sunglasses, heavy filters or a group shot. Use a clear, close photo and tell the model to keep the face from the reference image.
Most versions are 10 to 15 seconds. The one made here is about 12 seconds, vertical 9:16, with music.