What's the best AI model to use with this prompt?

This prompt works great with Sora 2, DALL·E 3, and Stable Diffusion. Any model that supports image-to-image transformations will produce fun results.

Can I tweak the style and quality arguments?

Absolutely! Change the `style` variable to things like "crayon" or "toddler drawing" and `quality` to "blurry" or "JPEG artifacts" for different comedic effects.

What kind of output should I expect?

Expect intentionally terrible, charmingly clumsy redraws that look like someone used a mouse in MS Paint for the first time. The worse it looks, the better it worked!

MS Paint Scribble Style Converter — Bad Drawing AI Prompt | sora2hub

GPT Image 2AI 提示词

MS Paint Scribble Style — AI Image Converter Prompt

Transform any image into a hilariously bad, pixelated MS Paint drawing with intentionally low-quality, clumsy aesthetics.

提示词 · EN

将附件中的图像重新绘制为 argstyle: 最笨拙、潦草且极其拙劣的风格。使用白色背景，使其看起来像是用鼠标在 MS Paint 中绘制的一样。它应该在视觉上似是而非，既有某种相似性，又在某种令人困惑且尴尬的层面上显得格格不入，并带有那种 argquality: 低质量的像素化质感，以突显其荒谬的拙劣感。说实话，无所谓了，你想怎么画就怎么画吧。

关于这个提示词

Transform any image into a hilariously bad, pixelated MS Paint drawing with intentionally low-quality, clumsy aesthetics. 它更适合作为 GPT Image 2 的概念艺术起点：先保留画面结构、主体关系和风格约束，再替换成你的品牌、人物或场景。

使用时优先替换 style、quality 等占位符，并保持镜头、构图和材质描述的顺序不变。这样更容易获得稳定的画面结果，也方便继续做多版本对比。