AI video prompt for a cinematic night creature video

Workflow

Copy this AI video prompt workflow for a WHAT LIVES BELOW night-creature clip. The page includes a real GPT Image keyframe and Seedance motion direction, plus sample output and reproduction notes for an image-to-video short.

GPT-Image2SeedDance 2.0

Pipeline

Step 1

分镜故事板

Image-to-imageGPT-Image2

生成方式: 参考生图

请基于上传的参考图 @image1、@image2，先生成符合下列设定的角色定稿关键帧/分镜静帧（Digital painting look，非写实 3D）：

SUBJECTS:
Emma @ image1, dark brown slightly wavy low ponytail, wearing a pink adventure jacket, off white T shirt, green cargo shorts, chunky sneakers, adventure backpack. Jake @ image2, dark brown fluffy short hair, wearing a green adventure jacket, off white T shirt, brown cargo shorts, chunky sneakers, adventure backpack.

ENVIRONMENT:
Nighttime tropical jungle ruins. A narrow moss covered stone path runs through the jungle. On the right is a large area of tall broadleaf plants and high grass, connected to a pitch black area with no visible interior behind the plants. On the left are tall damp stone pillars.

STYLE:
Digital painting look. Large simplified color blocks, hard edge brushwork, low detail structure, heavy shadows, visible hand painted texture. No realistic 3D look, no commercial CG ad look, no polished realism.