youtube_skill

This skill helps you search YouTube videos, fetch transcripts and channel details using YouTube Data API v3 with MCP server or yt-dlp fallback.

Python

2.6k

GitHub Stars

3

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

_meta.json445 B
README.md1.3 KB
SKILL.md7.3 KB

Overview

This skill lets you search YouTube, retrieve video and channel metadata, and fetch transcripts using the YouTube Data API v3 via an MCP server with a yt-dlp fallback. It consolidates search, channel listings, video details, playlists, and subtitle extraction into a single CLI-driven workflow for research, archiving, and content analysis. The design prioritizes API quota efficiency by using transcript lookups when possible.

How this skill works

The skill primarily calls a local MCP server that wraps the YouTube Data API v3 to perform searches, list channel videos, and fetch video metadata. For transcripts it first attempts MCP transcript extraction; if that fails or returns empty results, it falls back to yt-dlp to download auto-subtitles or extract VTT files. API keys are provided via Clawdbot config or an environment variable and all network calls go to Google APIs.

When to use it

Researching episode content or extracting quotes from podcast/video transcripts
Building a list of recent uploads from a channel for monitoring or archiving
Searching YouTube for topic discovery and collecting video metadata
Fetching captions when the official API returns no transcript
Preparing content summaries, social media posts, or research notes from videos

Best practices

Store the API key in Clawdbot config (~/.clawdbot/clawdbot.json) or YOUTUBE_API_KEY env var, not in scripts or repos
Restrict your API key to YouTube Data API v3 in Google Cloud Console to reduce risk
Try MCP transcript extraction first to avoid quota usage; use yt-dlp only as a fallback
Cache search and video detail responses when running repeated queries to save quota
Verify auto-generated transcripts before quoting—auto captions can contain errors

Example use cases

Find the latest episode from a podcast channel and pull its transcript for show notes
Search for videos on a trending topic, download transcripts, and analyze sentiment or claims
Collect video metadata and publish dates to build an archive or feed into analytics
Recover captions for videos missing API-accessible transcripts using yt-dlp VTT extraction
List recent channel uploads for monitoring competitors or creators you follow

FAQ

Set YOUTUBE_API_KEY in your environment or add the key to Clawdbot config, then rebuild or run the MCP server with the key present.

When should I use yt-dlp instead of the MCP server?

Use yt-dlp when MCP transcript extraction returns empty or when captions are unavailable via the API; yt-dlp can fetch auto-subtitles and produce VTT files.