---
name: video-repurposer
description: >
  Turn one long video or podcast transcript into a set of derivatives: timestamped
  chapters, short clip picks with start and end times, an article outline built from
  what was said, LinkedIn and X posts, and a newsletter blurb. Every quote stays
  verbatim with its timestamp. Use after a video, webinar, or podcast episode ships
  and the only thing published from it is the recording.
license: MIT
source: https://vibemarketers.ai/starter-pack/video-repurposer/SKILL.md
---

# video-repurposer

Get every useful piece out of one recording without putting words in the speaker's
mouth.

## When this applies

A long recording exists and nothing else was made from it. The ideas in it are good,
but nobody will watch the full hour. This skill cuts it into pieces people will
actually see, and keeps each piece traceable to the moment it came from.

## Inputs

- **A transcript with timestamps.** If there is only a recording, transcribe it first
  and keep word-level or sentence-level times.
- **The recording's URL**, so clips and posts can link back.
- **The marketing brain** if one exists, for voice and audience.
- **Channels wanted:** which of the outputs below to produce.

## What you produce

1. **Chapters.** Timestamp and a plain title for each topic shift. Titles say what is
   covered, not how exciting it is.
2. **Clip picks.** Three to six, each with:

```
start: 00:14:32   end: 00:15:41
hook line: "<the first sentence, verbatim>"
why it stands alone: <one line>
suggested caption: <one line>
```

3. **Article outline.** An answer-ready outline: the question the recording answers,
   the direct answer in the speaker's words, then sections built from the chapters.
   Each section cites the timestamps it draws from.
4. **Social posts.** LinkedIn and X drafts, each built around one idea and one
   verbatim quote with its timestamp, linking to the recording.
5. **Newsletter blurb.** A short paragraph: what the recording covers, the one idea
   worth the reader's time, and the link.

## Rules

- **Every quote is verbatim** from the transcript and carries its timestamp. Clean up
  filler words only, and never change meaning.
- **Nothing is claimed that the speaker did not say.** If a post needs context the
  recording lacks, flag it rather than adding it.
- **Each clip stands alone.** A viewer with no context must understand it. If a clip
  needs the previous minute to make sense, extend it or drop it.
- **No new numbers.** A figure appears only if the speaker said it, and it keeps the
  timestamp.
- **Do not publish.** Output is drafts for a human to approve.

## Self-check before you hand it over

1. Search the transcript for every quote. Does each one appear, at the stated time?
2. Watch or read each clip range cold. Does it make sense without the rest?
3. Does every post link to the recording?
4. Do chapter timestamps increase and cover the whole recording?