
Powerpuff Girls Duo Transformation
Image
GeminiPrompt Type: Image generation prompt
Click to try on Gemini
2
Number of original images required
1,086
Total number of views
Use two uploaded reference images (Image A and Image B).
Each image contains one subject.
Transform each subject into a Powerpuff Girls–style character, then place them side by side in a single shared composition.
Art Style & Proportions (STRICT)
Classic Powerpuff Girls aesthetic (original Cartoon Network style)
Very large round head (dominant part of the body, ~70 percent of height)
Huge round eyes taking up most of the face
Tiny dot mouth, minimal or no nose
Extremely small body with:
Short cylindrical torso
Tiny stick-like arms and legs
Flat 2D artwork
Bold black outlines
Bright, flat colors only
No gradients
No shadows
No 3D elements
Pure hand-drawn cartoon look
Character Design (Per Subject)
For each subject individually:
Use only their own uploaded image as the identity reference
Convert:
Hairstyle and hair color → simplified Powerpuff shapes
Skin tone → flat Powerpuff-style color
Match the exact outfit and accessories:
Clothing simplified and miniaturized
Accessories reduced to bold silhouettes
Preserve:
Hairstyle shape
Glasses shape (if any)
Outfit colors
Overall vibe and energy
Keep each subject recognizable despite exaggeration
Do not blend features between subjects
Pose & Composition
Both characters shown full body
Typical Powerpuff pose:
Slightly floating or hovering
Small feet off the ground
Place the two characters side by side, evenly spaced
Both characters facing forward
Balanced, symmetrical composition
No overlapping bodies
Background
Simple pastel or flat-color background
Background must not overpower the characters
Lighting
None
Flat cartoon lighting only
No shading, depth, or realism
Hard Rules (STRICT)
Do not alter either subject’s identity beyond Powerpuff stylization
Do not add extra characters
Do not add props or objects not present in the original photos
Do not mix or merge facial features between the two subjects
Each character must clearly map one-to-one to their own uploaded image