Plan the story
Zombie Love Story Prompt: A Clear Scene-by-Scene Template
A useful video prompt describes who is present, what changes, and how the camera follows the emotion. More adjectives do not automatically produce a more coherent clip.
Updated
Create your zombie videoAn original prompt to adapt
Create a short fictional cinematic scene using the two adult reference subjects. The first is the survivor; the second has a subtle, non-graphic zombie transformation. They recognize one another in a quiet abandoned street. Fear softens into recognition, and they share a gentle embrace. Transition to a warm daylight memory of them walking together. Keep their clothing and facial identities consistent. No injury, blood, attack, or on-screen text. Use restrained camera movement and a clear emotional progression.
Specify roles before style
Replace ambiguous words such as “they” with “the survivor” or “the transformed person” when an action could belong to either character. Describe one main action per beat and avoid asking for several simultaneous camera moves.
For a pet story, replace the second adult with the pet reference and describe an animal-appropriate interaction, such as the survivor kneeling beside it. Do not leave contradictory instructions about adult arms or human gestures in the prompt.
Adapt the prompt to your tool
Video providers differ in their supported duration, resolution, reference-image count, audio, and content rules. Keep the prompt within the settings offered by the tool you actually use. A text prompt cannot enable a capability that the model does not support.
Our studio supplies the scene instructions for its selected effect. This example is for planning or for use in a tool with a custom prompt field; it does not imply that every model will produce the exact sequence.
Before the first take
Should I put exact text or a caption in the prompt?
Usually add captions in an editor after export. Video models can distort letters, while a text layer in an editor gives you control over spelling, position, and timing.
Why does the ending sometimes differ?
A model interprets the prompt rather than executing a frame-by-frame script. Complex instructions, short duration, and unclear subject roles can make a sequence less consistent. Simplify the scene and inspect the result.