PROMPT RADAR / VIDEO视频叙事

个人资料 / 头像 - Anime Cafe Street Selfie

A polished anime-style prompt for generating a cheerful twin-tailed girl taking a sunny cafe-street selfie.

@TeshimaKairei 0 热度 0 浏览
个人资料 / 头像 - Anime Cafe Street SelfieORIGINAL VISUAL / 来源见文末
编辑先看

它为什么值得进入灵感库

它最有价值的部分,是把一个模糊想法推进到了可以被模型执行的程度。 把一段描述拆成起始状态、动作变化、镜头响应和结束画面,连续性比形容词更重要。

动作与镜头 Prompt

Create a bright, high-detail Japanese anime style selfie portrait of a cheerful young woman on a sunny European-style cafe street. The camera is held at arm’s length from a slightly high, wide-angle perspective, with her extended arm dominating the lower foreground and her face centered close to the viewer. She has {argument name="hair color" default="soft chocolate brown"} hair in long twin ponytails tied with black ribbon bows, wispy bangs, loose flowing strands, large sparkling {argument name="eye color" default="pink-purple"} eyes, rosy cheeks, and an open happy smile. Dress her in a feminine casual outfit: a white ribbed camisole with frilled straps and a small black bow at the neckline, a pale pink oversized cardigan slipping off one shoulder, a dusty pink plaid high-waisted skirt, a black belt with a gold buckle, delicate heart earrings, a small necklace, and a small black shoulder bag with a gold chain. Place her on a sunlit sidewalk lined with leafy green trees, a black streetlamp, storefront windows, flower planters, and a cafe with a pink awning; include one chalkboard sign on the left that reads {argument name="cafe sign text" default="Cafe de Floral"}. Use warm morning light, soft bokeh background pedestrians, dappled tree shadows on the pavement, glossy anime rendering, delicate linework, pastel colors, detailed hair highlights, and a romantic slice-of-life mood. Vertical portrait composition, 4:5 aspect ratio, crisp character focus with softly blurred background, no watermark, no extra text besides the cafe sign.
查看原作者内容

不要一次改完所有变量

用 Kling 复刻时,建议先单独生成稳定首帧,再把动作拆成“开始、变化、收束”三段;每段只保留一个主要运动。

  1. 主体:A polished anime-style prompt for generating a cheerful twin-tailed girl taking a sunny cafe-street selfie.
  2. 视觉方向:社区实测视觉提示词;按原提示词的光线描述执行
  3. 镜头与构图:按原提示词的镜头与景别执行;按原提示词的构图与画幅执行
  4. 交付约束:auto,遵循原提示词中的避免项;不要水印、乱码、低清、畸形手、错误透视或主体漂移。
负向约束

遵循原提示词中的避免项;不要水印、乱码、低清、畸形手、错误透视或主体漂移。

模型建议

Kling / Seedance / Veo。先用原始 Prompt 建立基准结果,再根据模型对自然语言、镜头运动或文字排版的能力做局部改写。

来源与编辑原则

保留作者,也保留判断。

本页收录公开案例并提供中文索引、结构化拆解和复刻建议,不冒充原作者作品,也不替代原帖上下文。

作者:手嶋海嶺(ゆっくり生命体) · 收录于 2026年8月14日

访问原始来源

沿着这条线继续看