Portrait direction library

Rumpelstiltskin AI Dance Prompt & Casting Ideas

The generator already owns the motion prompt. Use these briefs to plan the face, costume, silhouette and optional witness that you want the directed sequence to preserve.

Visible face
Readable expression
Costume shape
Role separation
Clean background

A strong casting image describes identity before the motion sequence begins.

Four original directions

Choose a clear visual identity

Copy a brief into your notes or use it as a checklist when taking a new portrait. These words do not get burned into the finished video.

Velvet stage magician

A sharply lit portrait of one adult performer in a dark velvet jacket, white formal shirt and expressive bow tie. Clear eyes, unobstructed face, restrained smile, strong silhouette, no crowd and no text.

Weathered forest traveler

One adult traveler with textured outerwear, windswept hair and a cautious sideways glance. Natural skin texture, visible jawline, balanced front light and a simple background.

Golden storybook creature

A friendly fantasy character with amber skin, curled ears, large readable eyes and a fitted green coat. Keep the face symmetrical, the costume edges crisp and both hands out of frame.

Formal guest and nervous witness

Photo 1: a confident adult in evening clothes with a playful expression. Photo 2: a different adult with wide, readable eyes and a worried reaction. Keep each portrait single-subject and front-facing.

Brief anatomy

What makes a portrait easier to direct

Identity survives motion more reliably when the important visual cues are visible before the first frame is generated.

Portrait choices and their effect on the dance
Decision	Include	Avoid
Face	Open eyes, visible cheeks and jaw, natural expression	Heavy shadow, sunglasses, cropped forehead or covered mouth
Hair and costume	Two or three distinctive, non-conflicting details	Long lists of tiny accessories that disappear during movement
Pose	Head-and-shoulders or waist-up, mostly facing camera	Extreme profile, crossed faces or another person touching the subject
Background	Plain or softly separated from the body	Crowds, mirrors, text, logos or objects cutting across the face

Two people, two distinct jobs

Keep the performer and witness visually separate

Photo 1: Defines the moving performer and must carry identity through body motion and closer facial shots.
Photo 2: Defines the seated witness only; a contrasting hairstyle or outfit makes the reaction easier to read.
One-photo version: Uses your portrait only for the performer and leaves the witness role supplied by the scene.