
原始文档
GPT Image 2 · en
Scene:
Same square eight-panel manga page format as the reference image, exactly eight panels in a 2 by 4 grid, same line art, screen tones, limited cyan, magenta, black, and cream palette, and polished product-launch energy.
Subject:
Use the same protagonist from the reference image with identical silver hair, round glasses, black cropped jacket, white hoodie, cyan scarf, and magenta sneakers. The same glowing glass orb with GPT-IMAGE-2 inside sits on a studio table, then floats, projects sketch thumbnails, and becomes the focus of a demo sequence across eight panels.
Important details:
Preserve protagonist identity and prop design from the reference. The studio has simple lights, a drawing table, camera tripod silhouettes, and clean product-demo staging. Keep speech bubbles as pseudo-text, with only GPT-IMAGE-2 readable on the orb. Each panel should advance the demo story while keeping the character model consistent.
Use case:
Third image in a reference-dependent comic consistency series, testing product demo storytelling with the same hero and prop.
Constraints:
PRESERVE the reference protagonist, outfit, orb, page style, panel count, and square crop. Exactly eight panels, no real logos, no watermark, no extra main character, no photorealistic UI, no readable dialogue except GPT-IMAGE-2, preserve exact 1:1 aspect ratio.使用方法
- 复制完整原始提示词,并保留其中的参数。
- 按需求替换主体、文字和构图要求。
- 在兼容的外部工具中使用,核对当前使用条件并检查结果。
开始前需了解
- 先确认是否需要参考图。标注兼容模型,不代表该示例由这个模型生成。
- 核对名称、数字与画面文字;不同模型版本的结果可能不同。





