Guide
AI Video Prompt Guide: How to Write Better Camera Movement Prompts
A complete guide to writing prompts that produce the camera moves you want in Seedance, Kling, Runway, Veo, Pika, and Luma.
The prompt formula
Subject + action + camera movement + style + lighting + pacing + aspect ratio
1. Subject — be specific
Not "a person running" but "a young female warrior in battle-worn armour sprinting through a collapsing corridor." Include character type, action, and environment.
2. Camera movement — the most important field
Use 1–3 movements. Connect them with "then". Common options:
- Dolly In — tension, focus, intimacy
- Pull Back — scale, isolation
- Orbit — hero energy, dimensionality
- Handheld — urgency, realism
- Crane Up — epic scale, finality
- Whip Pan — kinetic energy, transitions
3. Style
Name a recognisable visual style: cinematic anime, photoreal IMAX, gritty found footage, glossy commercial.
4. Lighting
Examples: golden-hour rim light, cold blue moonlight, pink and cyan neon, volumetric god rays.
5. Pacing
Slow, steady, fast, or frantic — affects how quickly the camera move plays out.
6. Aspect ratio
9:16 for TikTok/Douyin, 16:9 for YouTube, 1:1 for Instagram, 2.39:1 for cinematic scope.
Common mistakes
- Too many moves. Use 1–3. Connect with "then."
- Vague subject. Name the character, action, and setting.
- Missing lighting. Without it, models produce flat or random light.
- Conflicting directions. Dolly In + Dolly Out in one prompt confuses the model.
- Overlong prompts. Over 250 words and models ignore later details.