AI Video Storyboard Script Generator

AI video tools often fail when prompts are vague, actions are impossible to animate, or character consistency breaks down. This skill prepares digital-human video scripts by turning a content script and visual assets into model-ready storyboard prompts.

Problem

• A content script may define story, tone, duration, and dialogue, but lacks camera, action, lighting, and constraints.
• Visual assets may exist without being bound to shots using @image:, @video:, or @element:.
• Natural-language prompts can trigger deformation, motion artifacts, or style drift.
• Multi-shot outputs may have inconsistent character appearance, lighting, quality, aspect ratio, or timing.

How It Works

The skill organizes prompts with an S-A-C-S-C structure: subject, action, scene, style, and constraints. It extracts plot, style, duration, dialogue, and asset references, then splits the video into scenes. For each shot, it specifies the referenced subject, a continuous action sequence, shot size and camera movement, lighting and texture, and stability constraints. Actions favor slow, natural, connectable movements, such as “turn head → pause → smile,” while avoiding high-risk patterns like jumping, sprinting, or complex multi-character interactions. The output also considers platform formats such as 16:9 or 9:16, and checks single-shot duration, character consistency, and asset coverage.

Boundaries

It is intended for AI video prompt generation and digital-human performance scripts, not editing, rendering, or code generation. If asset lists or duration targets are missing, they need to be supplied. High-risk actions should be rewritten as more stable camera and motion designs.