原文
GPT Image 2 · en
Scene:
A square eight-panel product-demo comic page, exactly eight panels in a 2 by 4 grid, clean manga-inspired line art, crisp panel borders, soft screen tones, and a restrained black, cream, cyan, and orange palette.
Subject:
A consistent young inventor character demonstrates a floating image-making tablet across all eight panels. The character has curly dark hair, a yellow utility vest, black turtleneck, and orange headphones. The tablet projects small picture panels, then the inventor compares them, adjusts a dial, and presents a polished final image in the last panel.
Important details:
Strong character consistency across panels, clear commercial demo storytelling, coherent panel-to-panel action, readable gesture language, pseudo-dialogue bubbles, simple sound-effect shapes, no dependency on reference images. The page should feel like a launch storyboard that could be used in a product presentation.
Use case:
Independent comic single image, testing whether one prompt can produce a complete eight-panel product demonstration with consistent character and prop design.
Constraints:
Exactly eight panels, no real logos, no watermark, no readable text, no extra main characters, no broken panel grid, no photorealism, no cluttered UI, preserve exact 1:1 aspect ratio.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。
