如何撰写适合口播与演示的 YouTube 脚本。
围绕观众看完后能理解什么或做什么来构建脚本。用可见的实例支持解释,并为转场留出时间。
按任务需求调整流程。
确定要传授的内容、演示和最终要点,再围绕现有视觉素材撰写口播。分配时间展示结果和解释例外,并通过朗读排练检查脚本。
- 你需要提供什么
- 主题、提供的事实和视频形式。
- 你会得到什么
- 包含章节和制作提示的长视频脚本。
查看输入和结果。
演示用输入和输出 · 教学示例,并非 WebAct 实时运行结果
示例 个输入项
Write a short tutorial script showing duplicate-invoice checks. Example records: supplier North, invoice 104, USD 80 appears twice; supplier South also has invoice 104 for USD 80. Explain why supplier identity matters.
完整示例
On screen: three invoice rows. Narration: Two rows show invoice 104 for eighty dollars from North. A third row has the same number and amount, but its supplier is South. On screen: highlight supplier and invoice-number columns. Narration: Start with the supplier and invoice number together. The two North records are candidates for review. The South record should remain separate because invoice numbers can repeat across suppliers. On screen: label the North pair for review. Narration: Compare the original documents before removing anything. Matching fields identify a possible duplicate; they do not confirm which record should be changed.
将这份输入载入提示词,再复制到 WebAct 中尝试任务。实际结果可能与示例不同。
决策与问题排查。
教程的每一秒都要有旁白吗?
给观众留出观察操作或结果的时间。与连续讲解相比,停顿可能让演示更易跟上。
为什么脚本读起来不错,却超过计划时长?
将口播和屏幕操作一起计时。删减次要内容或拆分教程,不要强迫加快旁白。
使用自己的资料试一试。
将任务提示词中的示例换成你的材料。保留所需条件,再把任务复制到 WebAct。
自定义并复制任务 ↑