youtube-watcher

What This Skill Does

The youtube-watcher skill is a specialized agent capability designed to bridge the gap between multimedia content and text-based analysis. By leveraging the power of yt-dlp, this tool extracts closed captions and auto-generated subtitles from YouTube videos, converting visual/auditory information into structured, readable text. This allows your AI agent to process video content as easily as it would a local document, unlocking capabilities such as automated summarization, targeted information retrieval, and content-based reasoning without needing to 'watch' the video in real-time.

Installation

To integrate this skill into your OpenClaw environment, use the following command in your terminal:

clawhub install openclaw/skills/skills/michaelgathara/youtube-watcher

Ensure that you have yt-dlp installed on your system PATH, as the script depends on this external library to interface with YouTube's servers effectively. If you do not have yt-dlp, you can typically install it via pip install yt-dlp.

Use Cases

1. Research & Education: Quickly extract technical explanations or tutorials from long-form educational content to study key concepts without reviewing the entire video.
2. Content Summarization: Generate executive summaries or cliff notes for lengthy podcasts, interviews, or webinars for rapid consumption.
3. Information Retrieval: Locate specific data points, quotes, or timestamps mentioned in a video by programmatically searching the retrieved transcript text.
4. Accessibility: Convert video-only lectures into accessible, searchable text transcripts for team documentation or reference materials.

Example Prompts

1. "Can you watch this video https://www.youtube.com/watch?v=dQw4w9WgXcQ and give me a 5-bullet point summary of the main arguments?"
2. "Search the transcript for this video and tell me exactly what the presenter said about the new neural architecture updates."
3. "Is there a mention of cybersecurity best practices in this video? If so, extract those specific points for my notes."

Tips & Limitations

• Subtitles Required: The effectiveness of this skill is entirely dependent on the availability of captions. If a video does not have CC or auto-generated subtitles enabled, the script will return an error.
• Performance: Keep in mind that for extremely long videos, the transcript text can be very large. Ensure your agent's context window can handle the transcript length.
• Network Dependency: Because this skill fetches data from an external platform, an active internet connection is required, and access may be rate-limited by YouTube if multiple requests are made in rapid succession.