
原文
GPT Image 2 · en
Scene:
A square cooking-hack thumbnail for mobile feeds, using a split composition: sizzling pan close-up on the right and expressive chef face on the left. The background is warm cream with orange and green accents, plus a clear top category bar and a bold empty hook block overlapping the pan edge. The overall look should be bright, simple, and instantly edible.
Subject:
A charismatic home-cook creator in an apron reacts with delight while lifting chopsticks above a pan where steam and one dramatic cheese or sauce pull creates a single irresistible food cue. The ingredients should stay large and recognizable: one pan, one bright garnish, one utensil. The right-side hook area uses non-readable blocks and one badge-like circle to imply the platform template.
Important details:
Thumbnail logic comes first. Keep the cook's face large and expressive, keep the food cue singular and obvious, and avoid restaurant-menu clutter. Use strong warm contrast so the frame survives downscaling. This should look like a repeatable short-video cover system for kitchen hacks rather than a plated-food advertisement.
Use case:
Short Video Cover Pack single study 7/9, focused on split-screen cooking-cover logic with face-plus-food clarity.
Constraints:
No readable text, no watermark, no logos, preserve exact 1:1 aspect ratio, no crowded ingredient spread, no multiple dishes, no greasy mess, no fine-print labels or recipe cards.使い方
- パラメータを含む元のプロンプト全体をコピーします。
- 被写体、文字、構図を自分の要件に置き換えます。
- 対応する外部ツールで使用し、利用条件と結果を確認します。
開始前の確認
- 参考画像の要否を確認してください。互換モデルの表示は、そのモデルでの生成実績を意味しません。
- 名前、数値、画像内の文字を確認します。モデルの版で結果は変わります。





