怎麼填變量：[千層麵] 換成任何"有層次"的菜效果最好，例如"千層蛋糕""漢堡""壽司卷""冰粉"；[淺粉彩] 可換"淡薄荷""淺杏色"來配合菜的顏色。想要四宮格合集，就分四次生成後拼圖，或者直接寫"2×2 四格，分別是[菜 1][菜 2][菜 3][菜 4]"。示例圖是 2×2 合集：左上千層面像一摞彩色紙層，右上果仁蜜餅用綠色紙表現果仁層，左下泡芙塔用摺紙小球堆成錐形，右下青木瓜沙拉的絲用紙條表現、紅辣椒和番茄飛在上方，每格旁邊都有小字配料列表和軸測線稿。

常見問題與調整：

• 出來像真實食物：重複強調"所有部分都是紙做的，看得到摺痕和紙邊"。
• 標註太多太碎：寫"只保留左側配料列表和右側一張分層線稿"。
• 折法太簡單像兒童手工：加"摺痕密集、層數多、幾何精確"。
• 做豎版海報：畫幅改 3:4，標註放到底部。

適合：美食科普海報、餐飲品牌創意物料、設計作品集；不適合需要展示真實菜品外觀的點單圖。

英文原版

INPUT: [layered dish]

SYSTEM: Render the input as a master-folded origami and paper-craft architectural sculpture. Do not hardcode paper types unless inevitable. Infer the fold-geometry logic, structural load-bearing creases, layer-separation mechanics, color-blocking strategy, and the tactile grain direction of the materials.

SEMANTIC SOLVE: ORIGAMI_FOOD_AUTOPSY =
  (INFER(fold_geometry FROM mountain_valley_ratio + structural_crease_load + interlocking_tabs + negative_space_utilization) ::5) +
  (INFER(material_logic FROM paper_grain_direction + GSM_weight + surface_coating + edge_burnishing + adhesive_visibility) ::4) +
  (INFER(color_blocking FROM ingredient_color_mapping + contrast_hierarchy + gradient_paper_layering + visual_weight_distribution) ::4) +
  (INFER(hidden_craft FROM fold_sequence_complexity + tolerance_margins + tool_marks + humidity_warping_prevention) ::3) -
  (messy torn paper + realistic food textures + digital 3D paper shaders + cluttered craft desk + childish simple folds) ::-4

COMPOSITION: One central food item completely reconstructed as an intricate, multi-layered paper-craft sculpture. The "ingredients" are represented by distinct, beautifully colored paper layers, cut and folded with razor-sharp precision. Show the "exploded" floating layers hovering above the base, revealing the internal fold structure and interlocking tabs.

STYLE DNA: Intricate paper cutting art ::0.35 high-end origami photography ::0.25 flat-shaded C4D stylization ::0.20 crisp vector-like edges ::0.15 soft directional macro lighting ::0.05

OUTPUT: Soft off-white or pastel background, elegant minimalist typography, razor-sharp crisp edges, flawless flat-shaded colors with subtle paper grain textures, premium negative space.

NEGATIVE: no holograms, no bioluminescent glows, no VR/AR elements, no realistic food photography, no messy torn paper edges, no visible digital 3D artifacts, no cluttered backgrounds, no watermark.

改編自 @Gdgtify 發佈、EvoLinkAI/awesome-gpt-image-2-API-and-Prompts 收錄的提示詞，倉庫許可證 CC0 1.0。