Kling Zombie Video: Make the AI Zombie Trend in Kling 3.0

Kling is one of the first apps people search for after seeing the AI zombie trend on TikTok and Reels. It can make the whole zombie love story (the standoff, the hug and the match cut to a happy memory) with your own two photos. It won't do it from a template, though: you upload both photos, write the story as shots and usually re-roll a few times before both faces hold. This guide covers the model, the settings, a prompt and the real cost in credits.

Which Kling model to use

• VIDEO 3.0 Omni: the one for this trend. It takes up to 7 reference images (or 4 alongside a reference video), makes up to 15 seconds per generation at 720p or 1080p with native audio, and has Custom Multi-Shot, where you write each shot and its length. The zombie half and the memory fit in a single take.
• VIDEO 3.0 and 3.0 Turbo: for single shots. Good for one scene at a time, such as the hug alone. To tell the full story with them, make two clips and join them in an editor.
• Kling 4.0: not for everyone yet. Announced in late September 2026 with 30-second clips and up to 15 references. Only the Flash version is out, in a limited rollout, as of mid-October 2026.

Kling’s zombie effect is not the zombie trend

Kling’s effect library includes a zombie transformation: one photo in, a short clip of that person turning into a zombie out. It is fun, but it is not the trend. The trend needs two people and a story arc: a survivor who can’t shoot, an embrace and a hard cut to a warm memory. That takes two reference photos and a prompt, which is what the steps below do.

What you need

• Photo 1: the survivor holding the gun, usually you. Front-facing, face fully visible, good light.
• Photo 2: the one who turns. One person per photo; group shots make the model mix faces.
• Enough credits for three or four takes. That is how many most people need before both faces hold up in the zombie shots.

Step by step (3.0 Omni, one take)

1. Open Video, then choose VIDEO 3.0 Omni in the model picker.
2. Upload photo 1, then photo 2 as reference images. In the prompt they are @Image1 and @Image2, so the order matters.
3. Switch to Custom Multi-Shot and add five shots, using the lengths and text in the prompt below. If you’d rather use one prompt box, paste the whole prompt as it is.
4. Set 9:16 for TikTok, Reels and Shorts, and 720p. 1080p costs a third more and a phone feed won’t show the difference.
5. Leave native audio on if you want breathing, wind and laughter; turn it off to save credits and add the trending sound later.
6. Generate and compare takes. Keep the one where the infected face still looks like photo 2 and the cut lands exactly on the hug.

Photorealistic, cinematic short film. @Image1 is the survivor. @Image2 is the one who has turned. Keep both faces, hairstyles and builds exactly like the reference images in every shot. No blood, no gore, no text, no logos.
Shot 1 (3s): cold blue-grey light, an abandoned wooden cabin at dusk during a zombie apocalypse, handheld camera. @Image1 stands in a broken doorway aiming a rifle, hands shaking, breathing hard.
Shot 2 (3s): @Image2 slowly turns toward the camera, infected: pale grey skin, dark veins around clouded eyes, torn dusty clothes. @Image2 stops, recognising @Image1.
Shot 3 (2s): close-up of @Image1, tears running down. They lower the rifle and let it fall.
Shot 4 (3s): @Image1 steps in and hugs @Image2 tightly; @Image2's arms slowly rise to hug back.
Shot 5 (4s): hard match cut on the exact same hug pose into a sunlit wildflower meadow at golden hour, warm film colour, steady camera. Both are healthy and smiling in clean clothes; @Image2 lifts @Image1 and spins them around, both laughing.
Sound: wind and creaking wood, heavy breathing, then birdsong and soft laughter. No music, no dialogue.

Want a different memory? Change the meadow in shot 5 to a beach at sunset, a cozy cabin, your first kitchen or an autumn park, and keep the rest. For the dog or cat version, use the pet prompt on our free zombie love story prompt page and keep the same five-shot timing.

If you only have single shots (two clips)

1. Clip 1: shots 1 to 4 of the prompt, ending on “@Image1 hugs @Image2”.
2. Clip 2: “Opening on @Image1 and @Image2 already hugging” in the meadow, both healthy, then the spin.
3. Join them in CapCut (how to edit the cut there), trim each clip so the hug poses match, and cut on the hug. That cut is the moment people rewatch.

What it costs

Kling charges per second of video. Kling’s own guide lists VIDEO 3.0 Omni at 9 credits a second at 720p with native audio (6 without) and 12 a second at 1080p (8 without). A full 15-second take is:

Free accounts get 66 credits a day, enough for a short test of the faces but not a full take. The Standard plan is $6.99 for the first month with 660 credits, about four full 720p takes with audio. Prices were checked in October 2026; Kling changes them often, so check the cost shown on the Generate button before you render.

Kling-specific problems

• The faces swap after the cut. Name @Image1 and @Image2 in every shot, as the prompt does, and say which one turned. If it keeps happening, make each person an Element from two to four photos.
• The zombie looks too gory. Keep “no blood, no gore” at the top, and describe the look (grey skin, dark veins, clouded eyes) instead of words like “rotting” or “bloody”.
• The cut comes at the wrong moment. In Custom Multi-Shot the cut follows your shot boundaries, so put the hug at the end of shot 4 and the meadow at the start of shot 5.
• You used the zombie effect by mistake. Effects work from one photo and ignore your prompt. Start from the VIDEO 3.0 Omni model instead.

FAQ

Skip the prompt and the re-rolls

In Kling, a full 15-second take costs 90 to 180 credits, and most people need three or four takes before both faces hold. Our AI zombie video generator is built for this one video: upload two photos, pick the memory, and get the finished story (standoff, hug and match cut, about 25 seconds, 15 for pets) in 720p with no watermark. $6.90 per video, no subscription, no prompt, no clips to stitch. If a render fails, it is credited back automatically. For other apps, see the Higgsfield guide, the Google Flow guide or the PixVerse guide, or compare them all on the best zombie AI video generator.