A Simple Framework
A practical way to write AI image prompts is to follow this structure:
subject → state/action → scene → spatial relationships → style/quality → size
Or even more simply:
Generate an image of + subject + action + scene
For example: "Create an image of a cat napping in a sunbeam on a windowsill."
Key Principles
1. Define the subject clearly
Don't just say "a toy" or "a character." Say something specific like "character figure" or "collectible figure."
2. Add state and scene
Describe what the subject is doing and where it exists.
3. Use relationship words
Words like behind, in front of, next to, and on top of define composition — more useful than extra adjectives.
4. Add style and size
For example: realistic product photography, cinematic lighting, highly detailed, 16:9 aspect ratio.
5. Keep editing
Nano Banana works best through iteration. A well-structured prompt is much easier to refine.
6. Creativity is your friend
Nano Banana is great at surreal objects and unique scenes. Let your imagination run wild.
Case Study: Character Image → Collectible Display Scene
Prompt used:
turn this photo into a character figure. Behind it, place a box with the character's image printed on it, and a computer showing the Blender modeling process on its screen. In front of the box, add a round plastic base with the character figure standing on it. set the scene indoors if possible, 16:9 aspect ratioWhy it works:
- Subject — transforms a flat image into a physical figure
- State — figure standing on a round plastic base (feels like a real product)
- Scene — indoors pulls it into a believable environment
- Spatial relationships —
Behind it/In front ofcreates depth and composition - Size —
16:9 aspect ratiosets the output constraint
To improve further, add: realistic product photography, cinematic lighting, clean composition, highly detailed.
Before and After
Before: Original Character Image

After: Collectible Figure Display Scene

This comparison clearly shows how a single structured prompt can transform a simple character image into a complete product-style visual.
Final Takeaway
A good AI image prompt doesn't need to be long — it just needs structure:
- What is the subject
- What state is it in
- What scene does it belong to
- How are elements arranged
- What kind of final image do you want
That structure is what makes the result feel complete.
