MD This Page: Turn Any Webpage into LLM-Ready Markdown in One Click

By Prahlad Menon 4 min read

You’re researching something. You find a great article. You want to feed it to Claude or ChatGPT for summarization, Q&A, or analysis. But copy-pasting dumps a mess of formatting, navigation elements, and invisible junk into your prompt.

.MD This Page solves this with one click.

What It Does

Right-click any webpage → ”.MD this page” (or press Alt+M) → get clean, structured Markdown.

The extension uses Mozilla’s Readability library — the same tech behind Firefox’s Reader View — to isolate the main content and strip everything else: ads, navbars, cookie banners, script tags, tracking pixels.

What you get:

  • Headings preserved with proper hierarchy (#, ##, ###)
  • Lists, code blocks, and formatting intact
  • Images and links optionally included or removed
  • A clean document ready for your AI workflow

Why Markdown Matters for LLMs

Raw HTML is a nightmare for language models. A typical webpage includes:

  • Nested <div> soup
  • Inline styles and scripts
  • Navigation menus repeated across pages
  • Ad containers, social widgets, cookie notices

This burns context tokens without adding signal. A 5,000-word article might balloon to 50,000 tokens as HTML.

Markdown fixes this:

  • Token efficiency: Same content in far fewer tokens
  • Structural clarity: Headings and lists map directly to logical hierarchy
  • Better reasoning: Clean input → cleaner model output
  • Consistent parsing: No DOM quirks or broken nesting

Think of it as preprocessing your data before feeding it to the model.

Features That Matter

Customizable output: Toggle images, links, metadata, source URL, or generate a document structure map. Configure once in the options page and it syncs across browser profiles.

Quick actions: Right-click to instantly copy Markdown, copy as a prompt (with a system instruction wrapper), or download as .md — no preview tab needed.

One-click toolbar mode: Enable in settings to make the toolbar button immediately run your preferred action instead of opening the preview.

Open source: MIT licensed, code on GitHub. Works on Chrome, Firefox, Edge, and Brave.

Practical Use Cases

  • Research synthesis: Convert multiple articles, then ask Claude to compare arguments or extract key findings
  • Documentation extraction: Pull clean docs from any site into your knowledge base
  • RAG pipelines: Quickly capture web sources as properly formatted chunks for vector stores
  • Prompt engineering: Copy as prompt includes formatting hints for the model
  • Note-taking: Download .md files into Obsidian or your markdown editor of choice

How It Compares

vs. Copy-paste: No contest. Copy-paste brings formatting garbage.

vs. Reader Mode: Reader Mode renders cleaner HTML but doesn’t give you Markdown. You still get HTML formatting issues when pasting into chat interfaces.

vs. r.jina.ai: Jina’s Reader is an API service — great for automation but requires network calls and doesn’t work offline. This extension runs locally in your browser.

vs. MarkDownload: MarkDownload is a more general-purpose markdown clipper with more configuration options. MD This Page is optimized specifically for the LLM use case: fast, clean, AI-ready output.

Install

Chrome/Edge/Brave:

Firefox:

Or build from source:

git clone https://github.com/Ademking/MD-This-Page
cd MD-This-Page
npm install && npm run build

The Bigger Picture

We’re in an era where AI workflows are becoming as routine as browser workflows. Tools like this represent a shift: instead of websites adapting to serve AI-friendly content (like site-md), users are extracting clean formats themselves.

Both approaches matter. Server-side markdown (llms.txt, site-md) is better for sites that want to support AI agents. Client-side extraction (this extension) works everywhere, even on sites that will never add AI support.

For anyone doing serious work with LLMs, having a one-click “clean this page for AI” button is a small utility with outsized impact. Install it, bind it to your muscle memory, and stop wasting tokens on HTML cruft.

Built by Ademking. MIT licensed.