Make your pair the heart of Zombie AI

Prepare 2 photos, with one subject per slot. Photo 1 becomes the man holding the gun. Photo 2 becomes the woman in both her infected and healthy scenes. The reference is included, so no separate performance recording or location image is required. A collage cannot reliably communicate separate role assignments; use individual images and check their placement before generating.

Use an ordinary portrait. The instructions create the infected appearance in the dark scenes and use the same person’s healthy appearance in the memories. A heavily altered horror image makes the everyday identity harder to establish. Keep the face, hairstyle, and intended clothing visible in the input.

This preset follows one supplied sequence with two assigned roles. There is no alternate ending or additional cast selector. Choose Zombie AI when you want this confrontation and memory structure. For a different plot, start a separate request in AI Video Maker or bring your own reference to the main workspace.

No. The template already carries the role instructions and a configured Seedance 2.5 Edit model. The compact input panel keeps those technical controls out of the way. Choose your photos and resolution here. For custom wording or a different reference, use the main AI Genjutsu workspace; a restored draft keeps its previously saved inputs.

This preset follows its wide reference, which runs about 26.5 seconds. The resolution control changes the requested output quality, without choosing another performance or extending the scene. If a different length or composition is essential, start with another permitted source in the main workspace instead of trying to change this fixed routine.

The result uses the audio from the reference rather than a newly generated song. Page previews show examples; check the reference soundtrack on your completed result. After generation, check the completed video in My creations. If it reports an audio-processing issue, review the available recovery action before sharing.

Yes. Watching the original and preparing your photos does not start a generation. A submitted request requires sign-in and enough credits for the price displayed on Generate. Resolution and the request details affect that amount. A new attempt is another generation, so check the input and estimate before deciding to run it again.

Play the video through to the end and check the change from the dim confrontation to daylight memories and the final embrace. Look for changes to faces, clothes or fur, and overlapping limbs. A sharp thumbnail only represents one moment. AI editing can vary across frames, so keep a take that works as a complete sequence, including its soundtrack.

Use My creations in this workspace or open your video history. Available completed videos can be played and downloaded using the actions provided for your account. Replaying or downloading an existing result does not generate a new take. Check the full scene and audio before saving the version you intend to share.

Use pictures you have permission to edit, and consider the people or characters represented in the result. Access to a preset does not grant rights to the reference footage or soundtrack. Confirm the permissions for your intended publication and follow the destination platform’s rules for AI-created media. Community adaptations on this page are examples of the visual idea, not promises of an identical result.