PROMPT RADAR / VIDEO视频叙事
"Luxury Bags Collection with AI influencer " This influencer doesn’t exi
"Luxury Bags Collection with AI influencer " This influencer doesn’t exist. None of this is real. Every shot, every angle, every detail of the luxury bag you’re seeing was created with AI. This is what AI looks like now.
@AmControo 905 热度 4.6 万 浏览
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。
01 / 原始方法
动作与镜头 Prompt
"Luxury Bags Collection with AI influencer " This influencer doesn’t exist. None of this is real. Every shot, every angle, every detail of the luxury bag you’re seeing was created with AI. This is what AI looks like now. Not cartoons. Not obvious filters. But hyper-real, scroll-stopping humans built with advanced prompting, reference control, and visual consistency. No camera. No photoshoot. No real models. This is how creators are building AI influencers, luxury product showcases, viral content, and brand-ready visuals without the cost of traditional production. Tools used: • ChatGPT 2 – Character + product reference images • Seedance 2.0 – Video generation • CapCut – Editing
02 / 复刻路径
不要一次改完所有变量
用 Seedance 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。
- 主体："Luxury Bags Collection with AI influencer " This influencer doesn’t exist. None of this is real. Every shot, every angle, every detail of the luxury bag you’re seeing was created with AI. This is what AI looks like now.
- 视觉方向：实时 AI 视频案例；查看原帖媒体与作者说明
- 镜头与构图：按原帖镜头流程执行；查看原帖媒体与工作流
- 交付约束：auto，保留原作者方法与归属；复刻时避免水印、低清、畸变、主体漂移和无关文字。
负向约束
保留原作者方法与归属；复刻时避免水印、低清、畸变、主体漂移和无关文字。
模型建议
Seedance / Veo / Kling。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