Save a Web Page as Clean Markdown

Article copied cleanly via Copy outerHTML

Copying just the <article> element gives clean, readable Markdown with no nav or ad noise.

HTML in (from DevTools Copy outerHTML on <article>):
<article><h1>How DNS Works</h1><p>DNS resolves <strong>names</strong> to IPs.</p></article>

Markdown out:
# How DNS Works

DNS resolves **names** to IPs.

What happens when you paste the whole page

Include the analytics script and you get its source text in your notes. This is why copying a tighter element matters.

HTML in (whole page fragment):
<script>ga('send','pageview');</script>
<style>.ad{display:block}</style>
<p>Real content.</p>

Markdown out:
ga('send','pageview');

.ad{display:block}

Real content.

→ Re-copy just the <article> to avoid the leaked script/style text.

A data table from a reference page

Reference pages often have comparison tables. With a header row, they convert to a clean GFM table you can search in your notes.

HTML in:
<table><thead><tr><th>Port</th><th>Service</th></tr></thead>
<tbody><tr><td>443</td><td>HTTPS</td></tr></tbody></table>

Markdown out:
| Port | Service |
| --- | --- |
| 443 | HTTPS |

Add provenance to your saved note

The tool gives you the body Markdown; add a source header yourself so the archive is traceable.

After conversion, prepend in your vault:
> Source: https://example.com/how-dns-works
> Captured: 2026-06-13

# How DNS Works

DNS resolves **names** to IPs.

A video embed leaves a gap

Tutorial pages with embedded videos lose the embed. Capture the URL so you can re-link it in your notes.

HTML in:
<iframe src="https://player.vimeo.com/video/12345"></iframe>
<p>Watch the walkthrough above.</p>

Markdown out:
Watch the walkthrough above.

→ The iframe produced nothing. Add the URL by hand:
[Walkthrough video](https://player.vimeo.com/video/12345)

This is a converter, not a crawler. You must provide the HTML (View Source or Copy outerHTML). It never makes a network request to a page, which is also why it's fully private.

Modern SPAs build content client-side, so View Source returns an empty shell and the Markdown is nearly empty. Use DevTools → Copy outerHTML on the rendered element instead — that captures the live DOM.

There is no readability/article-extraction step. Whatever HTML you paste is converted. The clean-up happens at capture time: copy only the <article>/main element, not the whole page.

If your copied HTML includes <script> (Google Analytics, ad tags), the script source appears as text in the Markdown — Turndown has no rule to drop it. Likewise <style> CSS. Copy a tighter element to avoid them.

<iframe> embeds (YouTube, Vimeo, Google Maps) produce empty output. If the embed matters for your archive, copy its src URL from the source HTML and add a Markdown link by hand.

Sites that lazy-load images often keep the real URL in data-src and put a placeholder in src. Turndown reads src, so you may capture ![](placeholder.gif). Capture after images load, or copy the real URL from data-src manually.

Only DOM that exists at copy time is converted. For an infinite-scroll page, scroll to load the section you want before Copy outerHTML, or you'll archive a fraction of the content.

Layout-only tables (no <thead>/<th>) are left as raw HTML in the Markdown, since the GFM rule needs a header. Repair them with md-table-repair if you need clean output.

Full HTML pages with inline SVG and scripts are huge and can blow past the Free 500,000-character cap. The tool reports the count and limit. Copy just the article (which also gives cleaner output) or upgrade for the 5,000,000-character Pro limit.