web-fetcher

What problem does it solve?

Fetch any URL as clean Markdown. Use instead of WebFetch for JS-rendered pages, login-required platforms (Twitter/X, zhihu, reddit, weibo, xiaohongshu, bilibili, etc.), and complex pages. Routes known platforms through OpenCLI (browser login state), others through Jina Reader / defuddle.md / markdown.new. Invoke when the user provides a URL to read, extract, summarize, or convert to markdown.

Core Features & Use Cases

• Known platforms via OpenCLI (uses browser login state) for reliable rendering and access control.
• Generic URLs through a deterministic fallback chain: Jina Reader → defuddle.md → markdown.new.
• Use Case: convert research articles, blog posts, or documentation pages into Markdown for offline reading or integration into notes.

Quick Start

Fetch a URL and output the cleaned Markdown to stdout or save it to a file.