youtube-cli

name	youtube-cli
description	Search for videos, extract metadata, retrieve transcripts, and fetch comments from YouTube. Use when analyzing video content for research, summarizing long-form content, or gathering information from YouTube URLs.

What This Skill Does

This skill uses yt-dlp to fetch metadata (title, uploader, views, duration, likes, optional comments) and youtube-transcript-api to retrieve captions for a YouTube video. It also includes search functionality via yt-dlp.

Search Videos

1. Use the script skills/youtube-cli/scripts/search.py.
2. Run the command: python skills/youtube-cli/scripts/search.py "<query>" [flags]
3. Available flags: --limit <N>, --json, --output (-o).
4. By default, output is markdown. Use --json for raw JSON and --output <path> to save to file.

Inspect a Video

1. Use the script skills/youtube-cli/scripts/video.py.
2. Run either:
   • python skills/youtube-cli/scripts/video.py <video_id> [flags]
   • python skills/youtube-cli/scripts/video.py --url "<youtube_url>" [flags]
3. Available flags: --title, --uploader, --duration, --views, --likes, --transcript, --description, --comments, --json, --output (-o).
4. If no field flags are provided, defaults are: title, uploader, duration, views.

Steps

1. Use the script skills/youtube-cli/scripts/search.py via the terminal. Run python skills/youtube-cli/scripts/search.py --help for usage details.
2. Use the script skills/youtube-cli/scripts/video.py via the terminal. Run python skills/youtube-cli/scripts/video.py --help for usage details.
3. For long transcripts, descriptions, or comment threads, always use the --output flag to save the data to a file rather than printing to the terminal to avoid truncation.
4. If a video is part of a series or you need to process multiple videos, chain these calls as necessary.

Common Failures

• 400 Bad Request: Often caused by outdated yt-dlp or dependency issues.
• No transcript available: Some videos do not have captions, or the transcript service is disabled.
• Argument error: Use either positional video_id or --url; do not pass a URL as positional text with extra unsupported flags.
• Terminal Truncation: For long outputs, always use the --output flag.

Pro-Tips for Research

• When researching, first search for the topic using search.py to identify the most relevant videos.
• Use video.py on the top results to extract metadata and assess relevance before extracting full transcripts.
• If you need to analyze a large number of videos, script the interaction using the terminal to batch process them.