A Single Prompt Turns a Photo Into a Photoreal-and-Anime Dual Portrait
The prompt demonstrates how far a single, well-constrained text instruction can go in a modern image model — producing a specific multi-character composition with consistent likeness, distinct art styles, and controlled emotional tone, all without post-processing or inpainting.
The prompt uses an uploaded photo as a definitive likeness reference and produces a tall, high-end composition. One side renders the subject photorealistically; the other places a smaller anime version of the same person in an affectionate pose — clinging to an arm, making finger hearts, or gazing upward with bright eyes. The real-life figure wears a mock-grumpy expression that reads as fond rather than hostile.
A soft pastel backdrop and flattering light push the aesthetic toward Korean influencer content. The prompt explicitly forbids anatomical errors, extra limbs, visual clutter, text, watermarks, and any tone that feels cold or unsettling. The result is a polished image built from a single, carefully constrained generation request.
The prompt's structure treats the anime clone not as a separate subject but as an extension of the same identity, which reduces the risk of mismatched facial features.
Specifying an emotional dynamic between the two figures — overt affection versus playful annoyance — gives the model a narrative constraint that likely improves coherence across the composition.
Negative prompts here do more than suppress artifacts; they define the emotional floor of the image by forbidding hostility, coldness, and unsettling vibes.