---

name: youtube-transcript description: Fetch transcripts from YouTube videos. Use when user says "get transcript", "fetch transcript", "YouTube transcript", "pull captions", "what does this video say", or pastes a YouTube URL and wants the text content extracted. Also use when user wants to research or summarize a YouTube video.

YouTube Transcript Extractor

Fetch and return transcripts from YouTube videos using the bundled Python script.

When to Activate

• User pastes a YouTube URL and wants the transcript
• User asks to "get the transcript" or "fetch captions" for a video
• User wants to research, summarize, or extract learnings from a YouTube video
• User provides multiple YouTube URLs for batch transcript extraction

How It Works

Step 1: Extract the transcript

Run the script with the video URL or ID:

uv run .claude/skills/youtube-transcript/scripts/get_transcript.py "VIDEO_URL_OR_ID"

For timestamped output:

uv run .claude/skills/youtube-transcript/scripts/get_transcript.py "VIDEO_URL_OR_ID" --timestamps

Step 2: Handle the output

Supported URL formats:

• https://www.youtube.com/watch?v=VIDEO_ID
• https://youtu.be/VIDEO_ID
• https://www.youtube.com/embed/VIDEO_ID
• Raw 11-character video ID

Output rules:

• Never modify the returned transcript text. Present it exactly as returned.
• When timestamps are not used, arrange the raw text into coherent paragraphs by topic for readability.
• When timestamps are used, keep the [MM:SS] or [HH:MM:SS] format as-is.

Step 3: Save if requested

If the user asks to save the transcript:

• Default filename: {video_id}-transcript.txt
• Save to the location the user specifies, or to the current working directory

Multiple Videos

When given multiple URLs, run the script for each video sequentially. Present each transcript with a clear heading identifying the video.

Limitations

• Requires the video to have captions enabled (manual or auto-generated)
• Will not work on videos with captions disabled
• Auto-generated captions may contain transcription errors - flag this to the user
• Requires Python 3.10+ and uv installed locally