Labsco
MCP SERVER

HTML to Markdown MCP

by levz0r

Turn a web page into Markdown the model can actually read: fetch by URL or hand over HTML you already have, cap the length, and keep the result on disk if it is worth keeping.

Page Fetching & Content ExtractionVerified
Summary
The cheapest way to stop pasting HTML into a conversation.

A single call does the whole job — fetch, extract, convert — and maxLength is the parameter that decides whether a long page costs you a paragraph or a chapter of context. There is no crawling and no sign-in step, so a page that needs a session returns whatever an anonymous fetch of it contains.

What it is

A two-tool converter: one call fetches a URL or accepts raw HTML and returns Markdown, the other writes Markdown to a file.

What you get
  • A page fetched from a URL and converted to Markdown, or HTML you already hold converted in place
  • Output capped by maxLength, with page metadata included on request
  • The converted result saved to a file, either from the conversion call itself or separately
Requirements

No account and no key — a reachable URL, or the HTML itself.

Setup effort

One command — npx html-to-markdown-mcp