How to Make an AI Zombie Video with Your Photos

A practical photo-to-video guide: cast your characters, choose vertical framing, check credits, and troubleshoot faces, motion and sound.

Last updated: 2026-10-21

Start with the story you can actually generate

Choose Classic Trend for a prepared zombie reunion, or Custom Story to choose a relationship, scene, plot and ending. Custom Story offers reunion, rescue, escape or transformation, with a warm memory, escape together or infection twist ending. You select structured options rather than writing prompts. Watch the available examples and read their source notes before spending credits.

1. Choose your cast and assign the roles

  • Use JPEG, PNG or WebP images up to 8 MB each. Put one clearly visible subject in each photo.
  • Photo 1 is the survivor or pet owner. Photo 2 is the person or pet that transforms. Check this order before you generate.
  • Choose a front-facing or three-quarter view with the face unobstructed. Avoid sunglasses, heavy filters and tiny faces in group photos.
  • Use photos of consenting adults, or your own pet. Confirm permission in the studio.
  • For two people, use Swap roles if the photos are in the wrong slots. In pet mode, use a person first and a pet second.
  • For Custom Story → Friendship, upload 1–4 individual portraits. Put each friend in the human or zombie group to set their initial role. Use one person per photo; one portrait represents a solo character, while 2–4 portraits make a group story. These roles replace the two-photo survivor/zombie slots.

2. Choose the length and format

| Length | Credits | Best starting point | | ---------- | ------- | ---------------------------------------- | | 5 seconds | 200 | A short encounter and quick reunion | | 10 seconds | 300 | More time for the reveal and recognition | | 15 seconds | 500 | The fullest built-in story sequence |

These are reference costs for the direct pet-video workflow. Classic Trend offers 5s, 15s and, when available, 30s. Custom Story uses a 15-second storyboard. The studio shows the current credit quote for your selected mode; check it before confirming.

Choose 9:16 vertical for a portrait composition or 16:9 landscape for a wider scene. Only formats supported by the selected mode are offered. Keep faces near the center and leave room above and below for text you may add later.

The output tier is shown in the studio before generation and depends on the mode and active model. A resolution tier is not the same as an aspect ratio; exact pixel dimensions are displayed when known. The format preview illustrates cropping of a demo, not a prediction of your final clip.

3. Review, sign in and generate

Confirm your photos, roles, scene, story options, duration, format, audio and credit quote. Sign in to save the video to your account. For Classic and Custom Story, review the generated storyboard before continuing to video. If your balance is short, choose a pack on the pricing page, then return to the studio. Unspent pack credits stay in your balance. Drafts help you resume on the same device.

Generation progress shows submission, queuing, rendering and saving. Keep the page open when practical. If you leave, return to your account’s My videos page to check progress. Do not submit repeatedly just because the model is taking time.

4. Watch the whole clip before sharing

Check identity, role order, face continuity, hands, the selected ending and audio. Download the version you want to keep. Audio depends on the selected mode and configured model; read the audio label before generating. Music is included only when its option is available and selected. The studio does not copy a demo soundtrack or promise word-for-word dialogue or frame-for-frame reproduction.

Add your own captions in an editor if needed. Review the whole frame before publishing, and identify the scene as AI-generated fiction.

Troubleshooting

| Problem | What to try | | ----------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------- | | A face is hard to recognize | Use a sharper photo with a larger, unobstructed face. Keep the photo slots consistent. | | The roles are reversed | Check photo 1 and photo 2; swap before submitting a new generation. | | Motion feels rushed | Try a longer supported duration. A new generation uses a new credit quote. | | A format is unavailable | Choose an offered format. The current model controls support. | | A provider rejects a portrait | Follow the displayed error. Use an eligible photo or contact support; repeated identical submissions will not fix a provider restriction. | | Generation fails | Reserved credits return when failure is confirmed. Check status before submitting again. | | Saving is delayed | Use Retry saving in My videos. This retries storage, not paid generation. |

Photo privacy and deleting a draft

Photos are sent to the model service named beside upload. They are not automatically published in a gallery. Drafts expire 24 hours after the latest studio activity and are cleared when next opened. Use Clear photos on this device for immediate draft removal. Saved videos have no scheduled expiry and remain until you delete them from My videos. Provider copies have separate retention rules. Read the privacy and retention details.

Open the video studio

Need help?

Contact support with the task or order ID, the selected format and the error. You do not need to attach your original photos to a support request.