2.5k
GitHub Stars
2
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill openclaw/skills --skill youtube-watcher-7- _meta.json290 B
- SKILL.md1.8 KB
Overview
This skill fetches and reads transcripts from YouTube videos to enable fast summarization, question answering, and information extraction. It is designed to produce plain-text transcripts that can be summarized or searched programmatically. Use it when you need reliable access to a video's spoken content without watching the video.
How this skill works
The skill uses yt-dlp to download or extract subtitle tracks (closed captions or auto-generated subtitles) from a YouTube URL and outputs a plain-text transcript. It expects a local provider process (OpenClawProvider) to be available for integration in the environment. Once a transcript is produced you can run summarization, keyword search, or QA routines against the text.
When to use it
- Summarizing long videos into concise notes or bullet points.
- Answering specific questions about a video’s content without viewing it.
- Extracting quotes, timestamps, or topic sections from lectures or talks.
- Indexing spoken content for search or archival purposes.
- Preparing show notes or captions for accessibility and reuse.
Best practices
- Ensure yt-dlp is installed and available on the system PATH before running the script.
- Confirm OpenClawProvider is running when integrating the skill into automated pipelines.
- Verify that the target video has closed captions or auto-generated subtitles; otherwise the script will fail.
- Post-process transcripts to remove timestamps or noise before running summarization models.
- Split very long transcripts into chunks before feeding them to models with context-size limits.
Example use cases
- Run the transcript tool on a recorded conference talk to create a 3-5 sentence summary for email distribution.
- Extract and index Q&A segments from panel discussions so users can search by topic.
- Find and copy exact quotes from interviews by searching the downloaded transcript text.
- Produce study notes from recorded lectures by extracting key points and timestamps.
- Check a video transcript for mentions of specific product names or topics for compliance review.
FAQ
yt-dlp must be installed and available on PATH. The skill also requires OpenClawProvider to be available in the environment for integration.
What happens if a video has no subtitles?
If the video has neither closed captions nor auto-generated subtitles, the transcript extraction will fail and the script will return an error.
Which languages are supported?
The skill can extract any subtitle language provided on the video. Auto-generated subtitles depend on YouTube’s language support and accuracy varies by language.