ai-video-storyboard-expert

Write the prompt in temporal order: start state, subject identity, action, direction, camera behavior, environment, lighting, continuity, end state, then negative constraints. Keep one primary action per shot and test high-risk shots with short clips first.

Define each shot with an ID, objective, duration, aspect ratio, start state, action, screen direction, camera movement, continuity anchors, end state, and negative constraints. Plan the sequence architecture (hook, action, payoff, CTA) before writing individual prompts.

One-shot prompts fail when actions are too complex or constraints are missing. State "one continuous unbroken shot", prohibit cuts, teleportation, and scene replacement, define a single camera path, and reduce action complexity or split into multiple shots.

Define identity anchors for face, hair, wardrobe, and accessories, and use reference images where the platform supports them. Keep lighting stable, change one variable at a time, and never assume the model maintains identity across clips without verification.

Only if a video-generation tool is available in the environment; otherwise it produces storyboards, prompts, continuity sheets, and a testing procedure. It never claims a video was generated, rendered, or verified without actual tool access.

Avoid it for still-image prompts, brand strategy, copywriting, final editing, sound design, or legal rights clearance. Those tasks belong to dedicated image, copywriting, social strategy, or post-production specialists.