When I said forever
A tense reunion in a dark room, an embrace, and a return to warm beach memories.
TWO PEOPLE. TWO RECOGNIZABLE FACES.
中文Zombie couple video doesn't look like you?
Try the same studio as our homepage, or check what to change before your next render.
Choose a template
Swipe or scroll to browse. Click to select an available template.
Create your video
What should you change first?
First check whether the faces are wrong from the opening frame or change later in the video. If they start wrong, review each photo and its assigned role. If they change during movement, note the exact turn, overlap or scene transition. Change one input at a time before deciding whether a different workflow is worth trying. No tool on this page has been proven to preserve your two faces better than another.
Why doesn't the zombie couple video look like us?
Use these as checks, not a diagnosis of a model's internal behaviour. The same visible problem can have more than one cause.
Neither person looks right at the start
Compare each opening face with its photo. Check whether facial detail is readable, whether a filter changes features and whether you uploaded the intended person. Try changing only the less clear photo.
Check the photo guideThe people or story roles are mixed up
Check Person 1 and Person 2 against the selected template. In the current two-adult setup, Person 1 is the living survivor and Person 2 is the zombie companion. A prompt asks the model to keep that mapping; the result still needs checking.
Review the role mappingOne face changes after a turn or hug
Find the first frame where that person changes. Note whether the face turns away, is covered by the other person or changes at a cut. Keep that timestamp for your next comparison.
Review motion checksThe zombie look hides the person
Check facial shape and proportions separately from pale skin, cloudy eyes or lighting. Consider a different available scene if the styling makes recognition difficult. Styling changes alone do not prove that the face matches.
Use the review checklist
Which photos should we upload?
Choose a clear reference for each person. The selected template determines which input modes are available; photo-story presets require two separate photos.
- Two individual portraits
- Put the intended survivor in Person 1 and the intended zombie companion in Person 2. Confirm the selected template's labels before submitting. Separate inputs make the mapping explicit; they do not guarantee a closer match.
- One shared photo
- Use this only when the selected template offers shared-photo mode. Both adults must be clearly visible. The current shared-photo instructions identify the left adult as Person 1 and the right adult as Person 2.
- Check that each face is readable at normal viewing size, rather than tiny in a distant group shot.
- Prefer an unobstructed portrait with even lighting and a natural expression; Runway's reference guide recommends this for its image workflow.
- Compare the photo with the person's current appearance. Avoid selecting an image whose filter obscures the features you want to retain.
- Use photos you have permission to upload. Read the upload field for supported files and size limits.
How to use our two-person studio
The studio above is the homepage component, with the same templates, inputs and generation service. Selecting files or reading this guide does not submit a render.
Choose a two-adult scene
Use a couple scene for this task. Family and pet presets have different subjects and input requirements. The template preview shows a reference scene, not a personalized likeness test.
Check each photo and role
Upload the two adults in the intended fields. For the current couple templates, Person 1 is the survivor and Person 2 is the zombie companion. Clothing follows the template; face references come from the photos.
Review settings before generating
Check the current credit quote and available settings. Sign in to submit. This is a video-generation flow; the studio does not offer a separate face-approval image stage before a render.
Review the entire result
Check both people at the opening, during contact and after any transition. A new submission is a separate task with its own displayed credit cost. Do not assume a retry is free.
What if the faces change during movement?
Separate a wrong starting face from a face that changes later. In a workflow that accepts a reviewed starting image, check that image before animating it. Runway's documented approach is to begin with simple motion and add details incrementally.
- Record where each face first becomes hard to recognize: a turn, an overlap, an embrace or a scene change.
- When your tool allows it, test one simple action before combining several movements and camera changes.
- Change one setting or reference at a time. Record the change so a new result can be compared fairly.
- In our reference-template studio, the selected film guides the action. A story-detail field is not a frame-by-frame motion editor.
Should you use a template or a reference-based workflow?
Compare the controls you need. These are documented workflow differences, not a same-photo quality ranking. Third-party accounts, current availability and charges must be checked with each provider.
| Workflow | What you can inspect or control | What remains unproven |
|---|---|---|
| Our shared studio | Choose a reference scene and assign separate photo inputs, or use shared-photo mode where offered. Review the generated video. | No separate pre-render face-approval stage, and no measured claim that it preserves likeness better than the alternatives. |
| Runway image references → video | Its documentation describes combining reference images, iterating the still image and then sending an image to video. | A chance to inspect the still image is not proof that both faces will remain recognizable throughout the video. |
| Kling Element Library | Its official guide describes reusable character elements built from references and combining multiple elements. | Character-reference controls are not a tested result for your couple photos or this zombie story. |
| GoEnhance zombie couple template | Its product page describes a shared couple-photo template and automatic zombie-role selection. | The provider's likeness statements have not been independently tested here. Check its current preview and download terms. |
How to check whether the result looks like both of you
Use your own reference photos beside the video. This is a review method, not a completed case study or a measured similarity score.
- Opening
- Can you recognize each person independently? Check face shape and proportions, not just hairstyle or clothing.
- Movement and contact
- Pause around turns and the embrace. Check whether identities swap, features change or a face becomes unreadable.
- After the transition
- Check the same two people again in the final scene. A good opening does not establish consistency later.
- Before another attempt
- Save the model, template, settings, number of attempts and changed input. Judge both faces; do not report only the better one.
We have not published a controlled comparison using the same couple photos and story across these tools. No generated example, success rate or likeness score is presented here as test evidence.
Questions before you try again
Is there a zombie couple video app that definitely looks more like us?
This guide cannot establish that. It compares documented controls and suggests checks. A reliable comparison needs authorized photos of both people, comparable scenes, recorded settings and review of the full videos.
Are two separate photos always better than one couple photo?
No universal improvement is established here. Two fields make each person's reference explicit. Some templates accept a shared photo, while our photo-story presets require two separate identity photos.
Can I approve both faces before paying for a video here?
The shared studio does not have a separate still-image face-approval step. Photo thumbnails show your uploaded inputs; template previews show reference films. Neither is a personalized generated-face preview.
Can I choose who becomes the zombie?
In the current two-adult setup, Person 1 is the survivor and Person 2 is the zombie companion. Put the intended person in each field before submitting. A shared photo uses the left/right mapping described above. Generated role accuracy still needs review.
Will a longer prompt or higher resolution fix the faces?
Neither is a guarantee of identity accuracy. Check references and role mapping first. A higher-resolution output does not by itself establish that either person resembles the supplied photo.
Will retrying cost credits?
A new submission is a separate generation task with the cost shown in the studio. Do not assume another render is free; review the quote before submitting.
Does this page guarantee both faces will stay the same?
No. Reference inputs and identity instructions express the intended result. They do not guarantee facial resemblance, role accuracy or consistency through every scene.
Sources and what they support
Official pages checked on October 15, 2026. They document controls and provider statements; they are not our hands-on generation results.
- Runway — Creating with Gen-4 Image References
Reference-photo recommendations, image iteration and the move from a still image to video.
- Runway — Gen-4 Video Prompting Guide
Starting with simple motion and introducing changes incrementally, for the documented Runway workflow.
- Kling — Element Library User Guide
Reusable reference elements and multi-element creation; no independent quality comparison.
- GoEnhance — AI Zombie Couple Video
The provider's shared-photo template and automatic role assignment; no independently verified likeness claim.