mic-360 Markdown Negotiation
Request any HTML page as clean, structured markdown by including Accept: text/markdown in your request headers. This strips navigation, scripts, SVGs, forms, and other non-content elements, returning only the semantic page content.
Usage
Send a standard HTTP GET with the Accept header:
Example request:
GET https://bhaumicsingh.dev/blog/my-post HTTP/1.1
Accept: text/markdown
Response headers:
Content-Type: text/markdown; charset=utf-8
x-markdown-tokens: 1842
Vary: Accept
Content-Signal: ai-train=no, search=yes, ai-input=no
The x-markdown-tokens header provides an estimated token count (approximately 4 characters per token) so you can budget context window usage before reading the body.
What Gets Stripped
The converter removes these non-content elements:
<script>,<style>,<noscript>tags<nav>,<footer>elements<svg>,<iframe>,<video>,<audio>,<canvas>embeds<form>,<button>interactive elements- Elements with
aria-hidden="true"
When to Use
- Reading blog posts or project pages — get the content without layout markup
- Indexing the entire site — combine with
/sitemap.xmlto crawl all pages as markdown - Context building — use
x-markdown-tokensto estimate cost before fetching - Comparing with
/llms-full.txt— the LLMs context file gives a pre-built summary, while markdown negotiation gives you the live page content
Scope
Works on any route that returns text/html with HTTP 200. Static assets (images, JSON, XML) are unaffected. If the page returns a non-200 status, the original response passes through unchanged.
Authentication
No authentication required. All pages are publicly accessible.