PROMPT RADAR / VIDEO视频叙事

THIS IS FKING INSANE THIS 22 YEAR OLD REPORTEDLY MADE $10K THIS MONTH AN

THIS IS F**KING INSANE THIS 22 YEAR OLD REPORTEDLY MADE $10K THIS MONTH AND BARELY TOUCHES VIDEO EDITING SOFTWARE. She didn’t get better at keyframing. She stopped doing the manual work. Here’s the system she built: → Ch

@He1s_Sammy 878 热度 2.1 万 浏览
编辑先看

它为什么值得进入灵感库

先别急着照抄整段 Prompt:真正有用的是主体、空间和镜头之间的约束关系。 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

THIS IS F**KING INSANE THIS 22 YEAR OLD REPORTEDLY MADE $10K THIS MONTH AND BARELY TOUCHES VIDEO EDITING SOFTWARE. She didn’t get better at keyframing. She stopped doing the manual work. Here’s the system she built: → ChatGPT = the brain → Picsart Flow + Kling 3.0 = the animator ChatGPT studies what’s performing across viral kids’ content and turns those patterns into detailed visual prompts for her original character. She feeds the prompts into Picsart Flow, sets the render parameters, and generates polished 3D clips. All from a browser. Under 5 minutes per clip. She spends roughly 30 minutes a day managing the pipeline instead of animating frame by frame. And that’s the real shift. She’s not an animator anymore. She’s orchestrating an automated content engine. The reported economics: → $1.5K–$6K per 1M YouTube views → <5 minutes per render → 2–3 videos published daily → ~$10K/month reportedly generated The character concept is hers. The AI handles everything from prompt → animation → final render. While most creators are still learning complicated 3D software, she skipped the manual production stack entirely. She’s not selling animation skill. She’s selling volume.
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:THIS IS F**KING INSANE THIS 22 YEAR OLD REPORTEDLY MADE $10K THIS MONTH AND BARELY TOUCHES VIDEO EDITING SOFTWARE. She didn’t get better at keyframing. She stopped doing the manual work. Here’s the system she built: → Ch
  2. 视觉方向:实时 AI 视频案例;查看原帖媒体与作者说明
  3. 镜头与构图:按原帖镜头流程执行;查看原帖媒体与工作流
  4. 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

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:S✧ᜰ · 收录于 2026年8月27日

访问原始来源

沿着这条线继续看