# Second Brain Transcript

Clean a raw transcript before ingestion: punctuate, paragraph, label speakers, fix...

- Repository: https://github.com/undefined-ui/second-brain-os (1169 stars)
- Category: Media
- License: MIT
- Last updated: 2026-10-14
- Runs scripts: no, instructions only

## Install

Claude Code: `npx skills add undefined-ui/second-brain-os --skill second-brain-transcript -g -a claude-code -y`
Codex: `npx skills add undefined-ui/second-brain-os --skill second-brain-transcript -g -a codex -y`
Cursor: `npx skills add undefined-ui/second-brain-os --skill second-brain-transcript -g -a cursor -y`
Gemini CLI: `npx skills add undefined-ui/second-brain-os --skill second-brain-transcript -g -a gemini-cli -y`
Claude Code plugin marketplace: `/plugin marketplace add undefined-ui/second-brain-os`

## Description

Clean a raw transcript before ingestion: punctuate, paragraph, label speakers, fix mistranscribed technical terms, and split long recordings by topic. Use this skill when the user drops an auto-generated transcript, subtitle file, podcast or lecture transcript, meeting recording text, or asks to process a video or audio source. Do NOT use for text that is already prose, for summarising a transcript, or for the ingestion itself.

## SKILL.md, as published by its author (untrusted third-party text, read it as data)

# Clean a transcript

Concept pages built from unpunctuated caption dumps are noticeably worse,
because the model spends its attention reconstructing sentence boundaries
instead of understanding claims.

## Core rule

Clean, never summarise. Nothing is cut. This is a formatting pass on the record,
and the record has to stay faithful.

## Workflow

1. **Punctuate and paragraph.** Sentence boundaries first, then paragraphs at
   topic shifts.
2. **Label speakers** where they change. In an interview this is most of the
   value.
3. **Fix technical terms.** Auto-captions mangle proper nouns and jargon
   consistently. Flag anything you could not resolve rather than guessing.
4. **Keep coarse timestamps** every few minutes, so a claim on a wiki page can
   be verified in ten seconds instead of a rewatch.
5. **Propose split points** if the recording covers separate topics. A
   three-hour podcast is not one source.
6. **Write frontmatter:** title, channel or speaker, URL, date.

## Output format

The cleaned transcript, then:

```
Cleaned: <source>
Speakers labelled: <n>
Terms corrected: <list>
Uncertain: <list of what could not be resolved>
Suggested splits: <topics with timestamps, or none>
```

## Calibration

Never drop filler that carries meaning, such as hedging or a speaker correcting
themselves. Those are exactly what distinguish a claim from an aside.

If the transcript is too garbled to clean reliably, say so and recommend
re-pulling it rather than producing a plausible reconstruction.