yt-transcript

What this skill does

# YouTube Transcript Skill

Fetch transcripts from YouTube videos for analysis, summarization, or content processing.

## When This Skill Activates

This skill automatically activates when:
- User asks to get, download, or extract YouTube transcripts
- User wants to summarize or analyze YouTube video content
- User mentions video captions or subtitles
- User needs to process YouTube video text content

## Quick Start

### Prerequisites

The skill requires `youtube-transcript-api` to be installed:

```bash
pip install youtube-transcript-api
```

Or with uv:

```bash
uv pip install youtube-transcript-api
```

### Basic Usage

Fetch a single transcript (outputs to stdout):

```bash
python ${CLAUDE_PLUGIN_ROOT}/scripts/get_transcript.py "https://youtu.be/dQw4w9WgXcQ"
```

Save to file:

```bash
python ${CLAUDE_PLUGIN_ROOT}/scripts/get_transcript.py "https://youtu.be/dQw4w9WgXcQ" --output transcript.txt
```

Fetch multiple videos:

```bash
python ${CLAUDE_PLUGIN_ROOT}/scripts/get_transcript.py \
  "https://youtu.be/VIDEO_ID_1" \
  "https://youtu.be/VIDEO_ID_2" \
  --output transcripts.txt
```

Get transcript without timestamps:

```bash
python ${CLAUDE_PLUGIN_ROOT}/scripts/get_transcript.py "https://youtu.be/dQw4w9WgXcQ" --no-timestamps
```

## Supported URL Formats

The skill handles various YouTube URL formats:

- Short links: `https://youtu.be/dQw4w9WgXcQ`
- Standard URLs: `https://www.youtube.com/watch?v=dQw4w9WgXcQ`
- Embed URLs: `https://www.youtube.com/embed/dQw4w9WgXcQ`
- Video IDs only: `dQw4w9WgXcQ`
- URLs with parameters: `https://youtu.be/dQw4w9WgXcQ?si=abcd1234`

## Output Format

Default output includes timestamps:

```
# Transcript for https://youtu.be/dQw4w9WgXcQ (Video ID: dQw4w9WgXcQ)

0.00  First line of the transcript
2.50  Second line of the transcript
5.00  Third line of the transcript
```

Without `--no-timestamps` flag, plain text output:

```
# Transcript for https://youtu.be/dQw4w9WgXcQ (Video ID: dQw4w9WgXcQ)

First line of the transcript
Second line of the transcript
Third line of the transcript
```

## Common Workflows

### Analyze Video Content

```bash
# Fetch transcript
python ${CLAUDE_PLUGIN_ROOT}/scripts/get_transcript.py "https://youtu.be/VIDEO_ID" --output video.txt

# Now Claude can read and analyze the content
```

### Summarize Multiple Videos

```bash
# Fetch all transcripts to one file
python ${CLAUDE_PLUGIN_ROOT}/scripts/get_transcript.py \
  "https://youtu.be/VIDEO1" \
  "https://youtu.be/VIDEO2" \
  "https://youtu.be/VIDEO3" \
  --output all_transcripts.txt
```

### Extract Quotes or Key Points

```bash
# Get clean text without timestamps for easier processing
python ${CLAUDE_PLUGIN_ROOT}/scripts/get_transcript.py "https://youtu.be/VIDEO_ID" --no-timestamps > clean.txt
```

## Troubleshooting

### No Transcript Available

Some videos don't have transcripts available. The script will report:
```
Error: Could not fetch transcript for: <url>
```

This can happen when:
- The video has no captions/subtitles
- The video is private or restricted
- The transcript is disabled by the uploader

### Transcripts Not Auto-Generated

YouTube auto-generates transcripts for many videos, but not all. If a video has no manual captions and no auto-generated captions, the transcript won't be available.

### Rate Limiting

For bulk operations, you may encounter rate limiting. The script fetches one transcript at a time to minimize issues.

## Requirements

- Python 3.8+
- `youtube-transcript-api` package
- Network access to YouTube's transcript servers

## Script Options

```
usage: get_transcript.py [-h] [-o OUTPUT] [--no-timestamps] urls [urls ...]

positional arguments:
  urls                  YouTube video URL(s)

options:
  -h, --help            show this help message and exit
  -o OUTPUT, --output OUTPUT
                        Output file (default: stdout)
  --no-timestamps       Exclude timestamps from output
```

## Examples

### Get transcript for analysis

```bash
python ${CLAUDE_PLUGIN_ROOT}/scripts/get_transcript.py "https://www.youtube.com/watch?v=tgbNymZ7vqY" --output talk.txt
```

### Fetch clean text for summarization

```bash
python ${CLAUDE_PLUGIN_ROOT}/scripts/get_transcript.py "https://youtu.be/VIDEO_ID" --no-timestamps > summary_input.txt
```

### Batch process a playlist's videos

```bash
# Build URLs from a list of video IDs
for vid in VIDEO_ID1 VIDEO_ID2 VIDEO_ID3; do
  python ${CLAUDE_PLUGIN_ROOT}/scripts/get_transcript.py "https://youtu.be/${vid}" --output "transcript_${vid}.txt"
done
```