Technique · reference-sheet-control · multi-model-pipeline · character-consistency

Movement Sheet → Animation

Animate any character through 16 custom poses with one image and one video model.

A two-step pipeline that started going viral on X in April 2026. Step 1: ask GPT Image 2 to draw a 4×4 instructional grid of poses for your topic (dance, chores, fitness, anything). Step 2: feed that sheet to Seedance 2.0 with your character reference, and it animates the whole 16-panel sequence into a coherent video. The same prompt skeleton has been replicated by half a dozen creators with only the topic swapped — meaning it is a real, parameterizable template you can plug your own character into.

Try it →Workbench preview

The two-step pipeline

Each step shows the exact prompt creators used. Step 2 (Seedance 2.0) is intentionally blank where the original author didn't publish that half — we add the canonical community pattern with a clear note instead.

Step 1GPT Image 2· tagged· 23 lines

[STYLE]
Monochrome grayscale illustration, 3D-rendered character, clean instructional reference sheet, white background, comic-style cell grid layout, technical diagram aesthetic.

[LAYOUT]
4×4 grid layout with a total of 16 panels. Each panel is separated by thin black border lines. Cells are numbered from 1 to 16, with consistent panel sizes.

[CHARACTER]
image1 (the same character appears consistently in all panels)

[PANEL STRUCTURE – per cell]
Top-left: bold number badge + English title text
Center: full-body character pose illustration
Bottom-left: English description text (3–4 lines)
Overlay: directional arrows indicating movement

[ARROWS / MOTION INDICATORS]
Curved arrows, straight arrows, and circular rotation indicators placed around the character to show motion flow and direction.

[RENDERING STYLE]
Highly detailed 3D sculpted style, soft studio lighting, subtle shadows, no color, grayscale shading, clean linework, game concept art quality.

[NEGATIVE]
No background scenery, no color tones, no additional characters, no complex background.

The eight tagged sections ([VISUAL STYLE] / [GRID LAYOUT] / [CHARACTER] / [WARDROBE] / [PANEL STRUCTURE] / [MOTION INDICATORS] / [RENDER QUALITY] / [RESTRICTIONS]) are the canonical skeleton. Only the [SEQUENCE] section needs to change between topics.

Step 2Seedance 2.0· plain· 1 line

Character from Image 1 performs the dance based on the breakdown in Image 2. Maintain the character's exact appearance throughout. Follow the panel order from top-left to bottom-right, smooth transitions, beat-synced choreography. Each step ~1 second.

Canonical pattern verified by case @Ciri_ai (sheet_08). Most viral creators in this collection (oggii_0, john_my07, sharon7, vortex_promos) did not publish their exact Seedance prompt — the community converged on this short, directional shape. Confirmed by @Iancu_ai's experiment showing short Seedance prompts beat long ones. Source: github.com/EvoLinkAI/GPT-Image-2-Seedance2-Workflow Case 13.

The two-step pipeline

In the wild