LLM Markdown Converter
Paste a webpage's HTML and this tool strips out nav, headers, footers, ads, and scripts, then converts just the main content into clean, llms.txt-style Markdown. Nothing is sent to a server โ everything runs in your browser.
How to use it
๐ก Faster: the area-picker bookmarklet
Skip copying the whole page source โ grab just the part you want instead.
Drag this button to your browser's bookmarks bar. On the page you want to convert, click the bookmarklet, hover to highlight the element you want, then click it โ the HTML for that element and all of its child elements is copied to your clipboard. Paste it into the box below. (Press Esc to cancel picking.)
- Copy the page source: On the page you want to convert, press Ctrl+U (view source), then Ctrl+A / Ctrl+C to copy the full HTML. For React/Vue-rendered pages, open DevTools (F12) โ Elements, right-click the
<html>tag, and choose "Copy outerHTML" to grab the actual rendered DOM instead. - Add the URL (optional): Enter the source URL at the top to resolve relative links/images to absolute ones and add a source line to the output.
- Paste it: Paste the copied HTML into the left-hand box.
- Convert: Click "Convert to Markdown." The result appears on the right โ copy it or download it as a
.mdfile. - Add to llms.txt: The "llms.txt entry preview" box below auto-generates a one-line
- [Title](URL): descriptionentry you can paste straight into a site's/llms.txt.
This tool never fetches a URL directly. Browser JavaScript can't reach most external sites due to CORS, and this site doesn't run a proxy server. Pasting the HTML yourself guarantees the content never passes through any server โ it's processed entirely in your browser.
What is llms.txt?
llms.txt is a file placed at a site's root (/llms.txt), similar to robots.txt or sitemap.xml, that helps LLMs quickly understand a site's key content without wading through navigation and ads. It's plain Markdown: an # title, a > summary, and ## sections of links. It isn't a search-engine standard, but it's increasingly adopted by developer-tool and documentation sites. Converting a page here gives you a ready entry (title, description, link) to add to that site's llms.txt. See this site's own example at /llms.txt.
Frequently Asked Questions
Is the HTML I paste sent anywhere?
No. The pasted HTML is parsed and converted in your browser using DOMParser, and it's never sent to or stored on a server.
Why doesn't entering a URL fetch the page automatically?
Browser JavaScript can't fetch most external sites directly because of CORS. Working around that would require routing through a third-party proxy server, which would send the page's content through an outside server โ contradicting this site's "nothing sent to a server" principle. Pasting HTML avoids that entirely.
How does it decide what counts as "main content"?
It looks for common content containers first โ <main>, <article>, role="main", #content โ and falls back to the full <body> if none are found. <nav>, <header>, <footer>, <aside>, <script>, and <style> are always removed.
Are tables and code blocks converted too?
Yes. <table> becomes a Markdown table, and <pre><code> becomes a fenced code block (```), preserving the language tag when a language-* class is present.
What happens to relative image/link paths?
If you fill in the source URL, relative href/src values are resolved to absolute URLs against it. Leave it blank and the original attribute values are kept as-is.
Are images included in the result?
Not by default. <img> tags โ both regular URL images and data: base64 images โ are stripped from the output, since base64-encoded images especially bloat the Markdown and waste LLM tokens for no benefit. Turn on the "Include images" checkbox above the tool to keep them as  instead.