Write a two-photo zombie love story prompt
Written by Kadenly · Updated
Bind the survivor and partner to separate portraits, then connect the infected encounter to a healthy memory with one embrace.
Purpose and limitations
Zombie Love Story creates a new fictional scene from two adult portraits, not a replacement of actors in a source video. Faces, clothing, the embrace and transition timing can vary. It does not guarantee the exact original zombie clip, choreography, soundtrack or likeness.
Assign the two reference roles
Choose clear adult portraits with permission. State that image 1 is always the survivor and image 2 is the partner, infected in the opening and healthy in the memory.
Connect the story through the embrace
Describe a cold encounter, recognition and hesitation, then open arms and an embrace. Request a match cut during the embrace into a warm healthy memory. Keep it non-graphic and avoid extra people or conflicting actions.
Allow time, then check continuity
Start at 15 seconds, review the credit quote and generate. Check faces, healthy versus infected appearance, hands and the transition. Simplify one instruction if a beat is missing; exact timing and likeness can vary.
English prompt example
This is the editable English default prompt from the Zombie AI Video tool, not a verified sample result. Photo 1 is the survivor; photo 2 is the partner in the infected opening and healthy memory. Keep these assignments when changing the setting or ending.
Create an original 15-second cinematic zombie love story using two adult portrait references. @Image1 is the healthy survivor throughout. @Image2 is their visibly infected partner in the opening and their healthy self only in the final memory. In a cold abandoned shelter, the survivor recognizes their partner and hesitates. Begin with a clear close-up of @Image2 as a recognizable non-graphic zombie: ashen gray skin, clearly milky-white clouded eyes and faint dark veins around the temples. Keep @Image1 healthy with natural skin and eyes. The infected appearance must be visible before the embrace; do not show gore, wounds or injuries. The survivor chooses compassion and opens their arms. As they embrace, match cut into a warm golden meadow where both are healthy, smiling and holding each other; @Image2 now has natural skin and clear normal eyes, visibly contrasting with the infected opening. Keep the same two identities and the embrace continuous across the transition. Use readable faces, gentle camera movement and natural proportions. Do not swap identities, add people, show blood, fire weapons, add text or copy an existing film. Soft ambient sound; no dialogue required.Input and result checklist
- Upload one clear, front-facing adult portrait per slot, with permission from both people. Photo 1 is the survivor and photo 2 is the partner.
- Start at 15 seconds for the encounter, hesitation, embrace and memory. For a shorter clip, simplify the story to fewer actions instead of expecting every beat to fit.
- The Seedance 2.0 Mini option supports 4–15 seconds at 480p or 720p. Check the actual duration, ratio, audio setting and credit quote before submitting.
- Keep the infected partner in the opening and the same healthy partner in the memory. Describe one continuous embrace as the transition between settings.
- Review both faces, hands, physical contact, the infected-to-healthy change and the first and last frames. Listen to playback when audio is enabled.
- A task recorded as failed restores its deducted credits. A successfully generated video remains charged if its creative result does not meet your preference.
Common issues and adjustments
Faces change or the two identities mix
Use separate, evenly lit portraits without obstruction. Repeat the survivor and partner assignments clearly, reduce overlapping movements and keep both faces readable. Reference photos guide likeness but do not guarantee an exact match.
The embrace does not switch into a healthy memory
Keep one clear transition: the embrace continues from a cold shelter into a warm meadow, with both people healthy. Remove extra locations and actions, and use 15 seconds. A requested match cut is creative direction, not frame-accurate source-video editing.
The clip is silent or the sound differs from the prompt
Check that an audio-enabled option was selected and request simple ambience. Sound, dialogue and music can vary; this workflow does not preserve an original soundtrack. Review the returned video rather than assuming a requested audio track is present.
Sources and related policies
These are original preparation tips and prompt examples from our team. Model capabilities change; check current settings before generating. For input requirements and usage rules, see: