edit-video

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

Edit a video

剪辑视频

Keep each project in

video-projects/<slug>/

, preserving the source recording. Read the workspace guide and the project's brief or DESIGN.md. If a direction is missing, propose a style from

style-library/registry.json

based on the footage.

Inspect duration, streams, and source resolution with ffprobe. Copy the source into the project's assets or reference an explicitly provided local path.
Read
```
docs/TOOLS-AND-API-KEYS.md
```
before service setup. Reuse a matching word transcript. Otherwise use the student's chosen provider; Nate defaults to
```
node scripts/transcribe-elevenlabs.mjs <source>
```
. OpenAI Whisper and local Whisper are supported workflow alternatives after transcript normalization. Check available credentials or local dependencies. Explain uploads and costs before any service call whose authorization is still missing.
Read
```
../cut-silences/SKILL.md
```
. Produce an EDL and a retimed transcript, and render the silence pass when editing is authorized.
Read
```
../cut-mistakes/SKILL.md
```
. Inspect candidates in context and preserve intentional emphasis. Record the reviewed cuts. If no mistakes need cutting, carry forward the silence output. Match every transcript to its actual video.
Read
```
../video-storytelling/SKILL.md
```
for the visual arc and
```
../hyperframes-video-beats/SKILL.md
```
for overlays. Write a beat sheet with transcript anchors, one visual idea per beat, and deliberate callbacks.
Read
```
MOTION_PHILOSOPHY.md
```
,
```
../hyperframes/SKILL.md
```
, and relevant GSAP references before authoring. Use
```
../style-library/SKILL.md
```
to select or adapt a card. Localize every asset used by the composition.
Run preflight and HyperFrames lint from the project, plus the transcript-sync validator for anchored sub-compositions. Review Studio before the draft.
Render a draft, inspect encoded frames and transitions, and listen across cut boundaries. Check face framing, text legibility, black frames, and A/V sync. Resolve failures before the final render. Follow any review approvals already provided; do not repeat approval requests for the same authorized action.

Deliver the final MP4 path, edit decisions, retimed transcript, composition, and

VERIFY.md

describing what was checked and any remaining limitations. An automated detector proposes edits; editorial judgment remains necessary.

将每个项目保存在

video-projects/<slug>/

中，并保留源录制文件。阅读工作区指南以及项目 brief 或 DESIGN.md。如果缺少创作方向，则根据素材从

style-library/registry.json

中提出一种风格。

使用 ffprobe 检查时长、流和源分辨率。将源文件复制到项目的 assets 目录，或引用明确提供的本地路径。
在配置服务前阅读
```
docs/TOOLS-AND-API-KEYS.md
```
。复用与源视频匹配的逐词转录稿。否则，使用学生选择的提供商；Nate 默认执行
```
node scripts/transcribe-elevenlabs.mjs <source>
```
。在完成转录稿标准化后，也可以使用 OpenAI Whisper 或本地 Whisper 作为工作流替代方案。检查可用凭据或本地依赖。对于尚未获得授权的服务调用，应先说明文件上传和费用情况。
阅读
```
../cut-silences/SKILL.md
```
。生成 EDL 和重新校准时间戳的转录稿，并在剪辑获得授权后渲染静音裁剪版本。
阅读
```
../cut-mistakes/SKILL.md
```
。结合上下文检查候选剪切片段，保留有意进行的强调。记录已审核的剪切点。如果无需剪除错误，则沿用静音裁剪阶段的输出。确保每份转录稿都与其对应视频一致。
阅读
```
../video-storytelling/SKILL.md
```
以了解视觉叙事弧线，并阅读
```
../hyperframes-video-beats/SKILL.md
```
以了解叠加层设计。编写 beat sheet，其中包含转录稿锚点、每个 beat 一个视觉创意，以及刻意的回扣设计。
开始创作前，阅读
```
MOTION_PHILOSOPHY.md
```
、
```
../hyperframes/SKILL.md
```
和相关 GSAP 参考资料。使用
```
../style-library/SKILL.md
```
选择或改编卡片。对 composition 中使用的每一项素材进行本地化处理。
在项目中运行 preflight 和 HyperFrames lint；对于带锚点的子合成，还要运行 transcript-sync validator。在制作草稿前审阅 Studio。
渲染草稿，检查编码后的画面和转场，并试听剪切边界两侧的内容。检查人物面部构图、文字可读性、黑帧和 A/V 同步。在最终渲染前解决所有问题。遵循此前已给出的审阅批准；对同一项已授权操作不要重复发起审批请求。

交付最终 MP4 路径、剪辑决策、重新校准时间戳的转录稿、composition，以及

VERIFY.md

，其中需说明已检查的内容和仍存在的限制。自动检测器负责提出剪辑建议，但最终仍需人工进行编辑判断。