PROMPT RADAR / VIDEO视频叙事
Do you realize what the hell just happened? Fable 5 just flipped the ent
Do you realize what the hell just happened? Fable 5 just flipped the entire website-building game right in front of you. Fully animated, studio-looking sites, smooth scrolling, 3D globes, video backgrounds, all mobile re
@undefinedKi 1.7k 热度 3.7 万 浏览
编辑先看
它为什么值得进入灵感库
这条内容值得留下的,不是表面的风格词,而是它提供了一条可以继续实验的画面路径。 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
Do you realize what the hell just happened? Fable 5 just flipped the entire website-building game right in front of you. Fully animated, studio-looking sites, smooth scrolling, 3D globes, video backgrounds, all mobile responsive, built in one session by describing what you want. No agency, no code, no months of work. Here's the workflow, start to finish: 1. Find a look on Pinterest. Grab an image with the vibe you want (3D, animation, depth). 2. Turn it into a video. Use an AI image-to-video tool (Higgsfield, Kling) to animate it. Tell it to move but not zoom or reposition. 3. Build the hero with Fable 5. In Claude Code, describe the hero section, drop your video in as the background, set your colors. 4. Build section by section. Ask for the next sections in the same style. Once Fable has your fonts, colors, and layout locked, it reuses them and avoids generic AI-slop pages. 5. Refine with references. Screenshot a background, blur it in Figma to grab the average color, and tell Fable to match it. 6. Post it to grow. Screen-record the site, post the clip, and tag the tools you used. Companies repost work that shows off their platform. The people building real things with Fable now are going to be way ahead when everyone else catches up. Bookmark this.
02 / 复刻路径
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。
- 主体:Do you realize what the hell just happened? Fable 5 just flipped the entire website-building game right in front of you. Fully animated, studio-looking sites, smooth scrolling, 3D globes, video backgrounds, all mobile re
- 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
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