How to Make a Zombie Couple Video from Photos

A zombie couple video is built around recognition rather than an attack. Photos supply the identities; the story settings guide the scene.

This guide describes the Zombie Couple workflow. It is not a claim that every generation will preserve a face perfectly or reach each moment at an exact second.

1. Prepare the character references

Use photos you have permission to upload and get consent from every adult shown. Start with clear faces, even lighting and enough detail to recognize the eyes, hair and face shape. Avoid heavy filters and objects covering the face.

Use two separate portraits, one adult per image. The simple form uses a fixed survivor and partner assignment. Read the photo and role guide before choosing a tight or unusually wide crop.

2. Let the built-in story guide the scene

The simple generator supplies a quiet reunion setting and a warm-memory ending. You do not need to browse templates or provide a reference film. The right-hand panel offers a clearly labelled 21-second story example, loaded only when you choose Play example. The current generation request is 12 seconds at 720p without audio, so the example is not a promise of the same timing or result. It is not sent to the model.

During generation the panel shows progress. When your result is ready, it takes the example's place. If a task fails, its status stays visible and you can choose to view the example separately.

3. Confirm the roles before submitting

The survivor is photo 1 and the companion is photo 2. If you want the opposite assignment, exchange the two portrait uploads before submitting. The simple form does not offer shared-photo or pet controls.

The survivor and companion should stay the same people across both settings. In the memory, both return to their healthy appearance; a new scene should not introduce a third character. These are story instructions, not a guarantee that the model will follow every detail.

4. You do not need to write a prompt

The server builds the story instructions from the two assigned roles. There is no story-detail field in the simple form. The original prompt guide is optional reading or an example to adapt in another supported tool, not another step you must complete here.

5. Review duration, frame and credits

The simple form uses Seedance 2.0 Mini, 12 seconds and 720p with generated audio disabled. It offers portrait, landscape or square framing, not separate model, duration or resolution selectors. The requested length does not guarantee a precise cut or all four story beats.

Portrait suits a tall composition, landscape leaves more room around the characters, and square changes how much of their movement fits. Check the selected frame before submitting.

Read the live credit quote before you submit. For the current photo-only Standard / Mini request, 12 seconds at 720p is 72 credits. That is a site credit example, not a fixed dollar price or a promise of one usable film per attempt. Standard here is a model label, not a credit-pack name.

6. Follow the existing task

Sign in, confirm the photo permission statement and submit once. If a request is processing, check its status rather than immediately sending it again. My videos and the workspace can help you return to the task while signed in to the same account.

If the status is unclear, keep the task details and avoid repeated paid submissions. A newly submitted request may consume additional credits. Check the original task and your credit record; do not assume a page refresh creates a free retry or a cash refund.

7. Inspect the video before adding music

Watch once without distractions. Can you tell which character is the survivor? Do the faces stay recognizable? Does the embrace have separate bodies and plausible contact? Does the brighter ending read as a memory of the same two characters?

Then check the scene change. Similar framing, body positions and screen direction make the connection easier to understand. A slow dissolve or a completely new pose may make it feel like a different story.

The current story request is silent. Add licensed audio or captions in a separate editor if you want them, and review the rules of the platform where you plan to post.

Practical troubleshooting

What you notice	What to check before another attempt
The wrong adult changes	Read the upload labels; photo 1 stays human and photo 2 changes. Exchange uploads to reverse the roles
A face changes between scenes	Try a clearer reference and a simpler action; identity is still not guaranteed
Hands or bodies merge	Reduce complex movement and overlapping contact; inspect the whole embrace
The warmer ending is missing	The built-in request aims for a memory transition but cannot guarantee the story timing
The video is silent	This is the current story setting, not evidence that your speakers are broken
The task seems stuck	Check the existing task and credit record before submitting again
A download link no longer works	Check the saved task; a provider URL is not permanent storage and reopening it does not recreate the video

Download and keep your own copy

When the result is available, download while the provider URL works and confirm the saved file plays. Keep a copy before publishing. Labelling a clip as AI-generated and having consent from the people shown are separate from choosing music you have the right to use.