AI Video Prompt Formula for Beginners: Write Better Prompts in 2026
You just signed up for Sora, Kling, Runway, or Veo. You type “a cat playing piano” and get… something that vaguely resembles a cat-like blob near a piano-shaped object.
The problem isn’t the tool. It’s the prompt.
This guide teaches you a simple, universal formula for writing AI video prompts that actually work — no experience required.
The Formula
Every good AI video prompt follows this structure:
[Subject] + [Action] + [Camera] + [Lighting] + [Style]
1. Subject (who/what)
Describe what’s in the frame:
- People: “A young woman with dark hair in a red dress”
- Objects: “A matte black smartphone on a marble surface”
- Animals: “An orange tabby cat with green eyes”
2. Action (what’s happening)
Describe the motion:
- People: “walks slowly toward the camera”
- Objects: “rotates 360 degrees on a turntable”
- Animals: “leaps from a shelf to a countertop”
3. Camera (how it’s shot)
Describe the camera:
- Shot type: “close-up,” “medium shot,” “wide shot”
- Movement: “static,” “slow dolly forward,” “tracking left”
- Angle: “eye level,” “low angle,” “overhead”
4. Lighting (how it’s lit)
Describe the light:
- Direction: “from the left,” “from above,” “backlit”
- Quality: “soft,” “harsh,” “diffused”
- Color: “warm golden,” “cool blue,” “neutral”
5. Style (how it looks)
Describe the visual style:
- Genre: “cinematic,” “documentary,” “anime”
- Reference: “Wes Anderson symmetry,” “Blade Runner neon”
- Mood: “moody,” “bright,” “ethereal”
Examples
Beginner prompt (before)
“A dog running in a park”
Better prompt (after formula)
“A golden retriever runs through a sunlit park. Camera tracks alongside at dog height. Soft morning light from the left. Cinematic, warm tones, shallow depth of field.”
Beginner prompt (before)
“A woman drinking coffee”
Better prompt (after formula)
“A woman in her 30s holds a white ceramic cup and sips coffee. Medium close-up at eye level, camera slowly dollies in. Soft window light from the right, warm color temperature. Cozy lifestyle aesthetic, Kinfolk magazine style.”
Quick Reference
| Element | What to describe | Example phrases |
|---|---|---|
| Subject | Who/what, appearance | “A young man in a dark suit,” “A red sports car” |
| Action | Motion, movement | “walks toward camera,” “rotates slowly,” “explodes” |
| Camera | Shot type, movement, angle | “close-up, slow dolly forward, eye level” |
| Lighting | Direction, quality, color | “soft light from above, warm golden hour” |
| Style | Genre, reference, mood | “cinematic,” “anime,” “Wes Anderson” |
The Shortcut: Extract from Reference Videos
If writing prompts from scratch feels hard, there’s a faster way:
- Find a video that looks like what you want
- Copy the link
- Open socialtoprompt.com
- Paste → Extract Prompt
- Get a structured prompt instantly
Study the extracted prompt to learn how visual elements are described. Over time, you’ll internalize the formula and write better prompts on your own.
FAQ
How long should my prompt be?
30-100 words for most tools. Runway prefers shorter (30-60 words). Sora handles longer (60-150 words).
Do all AI video tools use the same prompt format?
Mostly. The formula works across all tools, but each has preferences:
- Runway: Short, motion-first
- Sora: Long, detailed
- Kling: Structured, four-part
- Veo: Cinematic terminology
- Seedance: Camera movement details
What if I don’t know anything about camera/lighting?
Use SocialToPrompt to extract prompts from videos you like. Study how the AI describes camera and lighting — then use those descriptions in your own prompts.
Conclusion
The formula is simple: Subject + Action + Camera + Lighting + Style. Master these five elements, and you’ll write prompts that produce stunning results on any AI video tool.
Shortcut: Extract prompts from reference videos at socialtoprompt.com — learn by example.
Last updated: September 2026
Share Article
