怎么填变量：上传一张角色清楚、最好是全身或半身的图片（原创角色、自己的插画或本人照片均可）。[2～3 头身] 可改成"1.5 头身"更圆更萌；[暖米色] 换成和角色主色相近的浅色。想要配套的图纸，可以在出图后追加一句："再生成一张它的纸模展开图：各部件平铺印在厚纸上，带虚线折线、粘贴边和编号，俯拍，旁边放剪刀和胶水。"

示例图是米色背景前的一个纸模小人：粉色的长卷发由一片片纸条叠成，戴一副棕色的大圆眼镜，围着粉色围巾，穿薄荷绿的毛衣和同色裤子、靴子，毛衣上是浅色的折线花纹；手脚是方块状的，能看到纸的折边，右下角有生成工具自带的四角星标记。

常见问题：

• 做成了光滑的黏土或塑料手办：强调"所有部件都是平面的纸折出来的，边缘是直线切口"。
• 不像原角色：把发色、发型和 2～3 个标志性配饰直接写进提示词。
• 纸片太碎太乱：写"部件数量适中，造型简洁"。

适合：角色周边和文创提案、社交头像、手工课和亲子活动的成品参考；用他人的角色或照片前请先取得许可。

日文原版

text
[AI Instruction: Image → SD Papercraft Conversion]

Analyze the [Input Image].

Extraction: Read the facial features, hairstyle, hair color, clothing, accessories, and pose of the person/character from the [Input Image].
SD Conversion: Convert into a Super Deformed (SD) character papercraft, 2-3 heads tall.
Output: Generate two images.
[Image 1: Unfolded Template/Parts]
The unfolded state of the papercraft. Colorful parts printed on thick paper. Unfolded templates for the head, face, hair, torso, arms, legs, accessories, etc. Fold lines (dotted lines), glue tabs, and numbers are visible. Scissors, cutters, glue, and a cutting mat are scattered in the background. Overhead view composition. Do not display any company names or logos.

[Image 2: Finished Product]
The assembled three-dimensional papercraft. SD character, 2-3 heads tall. Hairstyle, hair color, clothing, and accessories are reproduced based on the input image. Paper texture, folds, rounded corners, and a handmade warmth. Composition and lighting suitable for product promotion photography.

[Caution]
Do not include any real company names, brand logos, or trademarks.

[Atmosphere]
Quality similar to photos in craft magazines or handmade works. Soft studio lighting, clean background. High resolution.

改编自 @lovart_jp 发布、YouMind-OpenLab/awesome-nano-banana-pro-prompts 收录的提示词，许可证 CC BY 4.0，本站已翻译并改写（修改内容见 changes）；示例图来自同一条目。