AI Video Generation

ShotDirectorSim built a shot card and blocked generate until consent/disclosure was on.

Build a shot in layers:

• Subject and setting — who or what, and where
• Action — one clear movement
• Camera — framing and motion
• Look — medium, light, color, mood
• Timing — duration and pace
• Constraints — details to preserve or avoid

Create a 6-second 16:9 video.
Subject: a red paper boat in a shallow rain puddle.
Action: the boat drifts slowly past one yellow leaf.
Camera: low close-up, gentle tracking from left to right.
Look: overcast afternoon, realistic, soft reflections.
Timing: calm continuous motion, no cuts.
Avoid: people, text, logos, sudden camera shake.

For a sequence, make a tiny shot list first. Generate each shot separately, reuse the same character or object description, and change only one variable when continuity breaks. Save prompts, seeds, references, and settings when the tool supports them.

• Storyboards, concept clips, and background motion
• Product or lesson visuals that would be hard to film
• Testing a visual idea before a larger production

Do not clone a real person’s appearance or create deceptive footage without permission and clear labeling. Check music, image, character, and commercial-use rights. Generated video can distort bodies, physics, text, and brand details frame to frame.

Write one six-second shot with all seven layers. Generate a second version that changes only camera movement, then compare which better serves the scene.