Seedance 2.5 Killer Prompt: High-Sakuga 2D Animation Fight Scene Tips
A practical guide to structuring prompts for dynamic camera moves and high-impact 2D anime fight scenes using the Seedance 2.5 AI video model.
Creator Kōda (@aimikoda) has shared a prompt structure and direction guide for generating high-sakuga 2D anime-style combat scenes with precise impact and dynamic camera movement using the Seedance 2.5 AI video generation model. This breakdown outlines practical techniques for overcoming static AI video limitations and maintaining continuous action momentum.
Image source: @aimikoda
The newly released guide goes beyond basic image generation, establishing concrete frameworks for character appearance specifications, three-dimensional camera trajectories, and visual feedback during impacts.
Character References and Visual Style Definition
To elicit accurate action sequences, distinct character specifications and overall visual styles must be firmly anchored.
- Character Specifications (
References):@[char1 ref]: Mina. Long black hair, green striped crop top, baggy orange cargo pants with black graffiti, white sneakers, black fingerless boxing gloves on both hands, chain bracelet.@[char2 ref]: Rue. Long pink hair, black crop top, black cargo shorts with chains and tags, bare knuckles with silver rings, skin visible on both hands, chunky lime sneakers, dog tag necklace.
- Visual Style (
Visual style): High-sakuga flat-color 2D anime with bold linework and hard cel shading. Flat blue and teal backgrounds that flash hot pink or orange on impact, speed smears, impact frames, and deep foreshortening.
Continuous Camera Movement and Impact Directing
The camera remains in constant motion, spiraling around crossing limbs, diving under kicks, and pulling viewers directly into the choreography.
- Action Energy (
Energy): Within every shot the camera keeps moving, skimming within inches of a fist or a sole. Every landed hit plays as a two-frame impact frame followed by long smearing and elastic recoil. - Negative Constraints (
Avoid): No subtitles. No slow motion, no idle posing, no pauses between actions, no freeze frames, or held two-shots. - Setting: A relentless brawl on an empty concrete rooftop with a chain-link fence and a steel vent box, driven entirely by hits, shoe scuffs, chain rattles, and wind without dialogue across a 30-second sequence.