怎麼填變量：[4 個世界奇蹟] 可以換成"4 座中國古建築（天壇、黃鶴樓、布達拉宮、應縣木塔）""4 座城市地標"；只做一個時，把第一句改成"單張 16:9"，並寫明 [地標建築] 和 [與地標相配的口味]，例如"[天壇]＋[抹茶、紅豆、桂花]"。

示例圖是四宮格：泰姬陵（白色蛋糕穹頂，粉綠開心果切片）、佩特拉古城（紅色巖壁蛋糕，覆盆子切片）、奇琴伊察金字塔（巧克力階梯蛋糕）、馬丘比丘（抹茶綠山形蛋糕），每格前景一塊分層切片在金勺上，左右兩側是配料小圓圖和口味標註小字。

常見問題：

• 地標不像：寫出地標的關鍵造型（"白色洋蔥形穹頂 + 四座尖塔"）。
• 切片層次亂：寫"切片可見 4～5 層，層次分明"。
• 標註小字亂碼：標註只是裝飾，想要可讀就減少到每格 3 條。

適合：甜品店創意內容、旅行主題社媒、美食帳號趣味圖、烘焙課程靈感板。

英文原版

text
2x2 grid, 16:9, do this for 4 wonders of the world, ai picks the rest for flavor profile ..: {   "Scene_Type": "Macro Food Photography with Graphic Design Overlay",   "Subject_Cake": "{argument name="landmark" default="[ARCHITECTURAL_LANDMARK]"} engineered as an Entremet",   "Flavor_Base": "{argument name="flavor" default="[FLAVOR_PROFILE]"}",   "Composition": {     "Background": "The main {argument name="landmark" default="[ARCHITECTURAL_LANDMARK]"} cake, slightly out of focus (bokeh).",     "Foreground": "A single wedge/slice pulled forward on a gold dessert spoon, in hyper-sharp focus.",     "Textures": "AI_INFER(High contrast between glossy mirror glazes, porous aerated sponge, and crumbly sablé base matching the flavor)."   },   "Infographic_Data": {     "Macro_Callouts": "Thin vector lines pointing directly to the microscopic crumbs and bubbles on the slice.",     "Data_Widgets": "AI_INFER(Small circular 'Micro View' inserts showing extreme close-ups of the ingredients)."   } }

改編自 Gadgetify 發佈、YouMind-OpenLab/awesome-gpt-image-2 收錄的提示詞，許可證 CC BY 4.0，本站已翻譯並改寫（修改內容見 changes）；示例圖來自同一條目。