Stats
Actions
Tags
From fetch
Fetches web content using MCP tools in HTML, Markdown, JSON, plain text, readable article, or YouTube transcript formats. Use as fallback when WebFetch fails with 403/access errors or for targeted extraction.
How this skill is triggered — by the user, by Claude, or both
Slash command
/fetch:fetchThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use fetch MCP tools to retrieve web content in the most appropriate format for the task.
Use fetch MCP tools to retrieve web content in the most appropriate format for the task.
| Tool | Best For |
|---|---|
mcp__fetch__fetch_markdown | General web pages — converts to clean Markdown |
mcp__fetch__fetch_html | Raw HTML when you need to parse structure |
mcp__fetch__fetch_txt | Plain text extraction, minimal formatting |
mcp__fetch__fetch_json | JSON REST APIs and data endpoints |
mcp__fetch__fetch_readable | Articles/blog posts — removes ads and navigation |
mcp__fetch__fetch_youtube_transcript | YouTube video transcripts by video URL or ID |
fetch_markdown or fetch_readablefetch_jsonfetch_htmlfetch_youtube_transcriptfetch_txtWhen WebFetch fails, the fetch plugin automatically suggests retrying with these tools. Prefer fetch_markdown as the general-purpose fallback for most URLs.
npx claudepluginhub pleaseai/claude-code-plugins --plugin fetchFetches any URL and returns clean Markdown via local trafilatura, with Exa MCP fallback for JS-rendered or anti-bot pages. Use instead of built-in WebFetch for reading, scraping, or summarizing web pages.
Fetches web content via curl and WebFetch for HTTP GET/POST, JSON APIs, HTML, auth, and cookies. Use when a specific URL is provided.
Fetches any URL as clean markdown, using OpenCLI for login-walled platforms (Twitter, zhihu, reddit, etc.) and Jina Reader/defuddle.md for generic pages.