Why this structure is useful
Consistency improves when the prompt repeats the few traits that matter most: name, silhouette, wardrobe, palette, location, and one action.
Copyable template
Feature [character name], defined by [age range/silhouette], [hair or facial marker], [wardrobe], and [color palette]. Place the character in [location] completing [one action]. Start in [opening frame], cut to [detail], and end on [closing frame]. Camera: [camera approach]. Preserve [identity and geography constraints] throughout.Variables to change
- name and silhouette
- facial marker and wardrobe
- palette and location
- one action and shot sequence
- identity and geography constraints
Three-shot plan
- Shot 1: Establish the character with the key identity markers visible.
- Shot 2: Show one simple action without changing wardrobe or proportions.
- Shot 3: Resolve on a recognizable character frame in the same scene geography.
Failure modes to avoid
- Adding many wardrobe, age, or mood changes creates competing identity instructions.
- Rapid angle changes without a stable scene anchor make character and location continuity harder to judge.
Continue planning
Open this template in the prompt builder
Review the Seedance 2.5 status tracker before relying on model access, pricing, or API claims.