From web-x
Extract content from web pages, video platforms, and multi-page sites. Routes to the right backend automatically — trafilatura for static pages, yt-dlp for video transcripts, crawl4ai for JS-rendered or deep crawls. Use when the user provides a URL to extract content from, wants a video transcript, or needs to crawl a site. Do NOT use for URLs ending in .md (use WebFetch). For simple article reads, the defuddle skill is a lighter alternative.
How this skill is triggered — by the user, by Claude, or both
Slash command
/web-x:webThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Unified web content extraction with three backends. Choose based on the source.
Unified web content extraction with three backends. Choose based on the source.
URL provided
├── Video platform? (youtube.com, youtu.be, vimeo.com, twitter.com/*/video,
│ tiktok.com, twitch.tv, soundcloud.com, dailymotion.com, etc.)
│ └── yt-dlp → transcript + metadata
│ Recipe: just web-transcript <url>
│
├── Single web page?
│ ├── Try trafilatura first (fast, no browser needed)
│ │ Recipe: just web-fetch <url>
│ ├── Result empty/thin? Likely JS-rendered (SPA, React, Angular, dashboard)
│ │ └── Escalate to crawl4ai
│ │ Recipe: just web-fetch <url> --js
│ └── Result good? → done
│
└── Multi-page site / deep crawl needed?
└── crawl4ai deep crawl
Recipe: just web-crawl <url> [depth]
just web-fetch <url>just web-transcript <url>just web-fetch <url> --js (single page) or just web-crawl <url> [depth] (multi-page)| Tool | When to prefer it |
|---|---|
| defuddle (obsidian plugin) | Quick article read already in conversation flow, Node-based |
| WebFetch (built-in) | URLs ending in .md, or when you need a model-summarized version |
| web-x:fetch (this plugin) | Raw markdown needed, full metadata, or defuddle/WebFetch return thin content |
| web-x:transcript (this plugin) | Any video URL — replaces disabled youtube_transcript MCP |
| web-x:crawl (this plugin) | Multi-page extraction, site-wide content gathering |
If a backend is not installed, recipes print a clear error with the install command:
Error: trafilatura not found. Install with: uv tool install trafilatura
Error: yt-dlp not found. Install with: uv tool install yt-dlp
Error: crawl4ai not found. Install with: uv tool install crawl4ai && crawl4ai-setup
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub rugved-rakebma/web-farm-market --plugin web-extract