What breaks when you save a web page as Markdown, and what to do. Updated July 2026.
Saving an article as Markdown mostly works, right up until it does not: the images are dead links, the code blocks lost their formatting, half the page is cookie banner, or you got three paragraphs of a ten-paragraph piece. Each of those has a different cause, and the fix is usually a different tool rather than a different setting.
Diagnosis
The clipper saved the page's image URLs rather than the images. It reads fine today and breaks the day the site reorganises or goes down. If you are clipping for an archive rather than for reading this week, you need a tool that downloads images alongside the text and rewrites the links to point at them.
Almost always lazy loading: the rest of the page had not rendered when the clipper read the DOM. Scroll to the bottom first, wait for it to settle, then clip. If a site does it aggressively, no clipper will beat it reliably.
The clipper is saving the raw DOM instead of running a readability pass. Tools built on Mozilla's Readability engine — the same one behind Firefox Reader View — strip that automatically. If yours does not, dismiss the banner and switch the browser to reader mode before clipping.
Syntax-highlighted code on the web is a nest of <span> elements, and a naive HTML-to-Markdown pass flattens it. This is where clippers differ most in quality. If you clip documentation, test one code-heavy page before settling on a tool.
Nothing on this page changes that, and a clipper that promised otherwise would be selling you something else. If you can read it in the browser, a clipper can usually save what you can see.
Choosing
€10, once. Lives in the menu bar, saves the current Safari page as an ordinary Markdown file in an ordinary folder — no vault, no extension, no account — and downloads the images next to it with the links rewritten. macOS 14+.
Free and open source, official, and built around templates so the front matter comes out the way your vault expects. If Obsidian is where the clip is going, this is the natural fit. If it is not, you are installing a vault to use a clipper.
Free and open source, one click, with templates for front matter and options for how images are handled. One caveat worth checking first: its repository has had no commits since June 2025. It still works for many people and a Manifest V3 fork exists, but an extension nobody is updating is a risk as browsers keep changing.
Free and open source, and not a Markdown tool at all, which is the point. It saves a faithful copy of the whole page as one HTML file with styling, fonts and images embedded, so it still looks right years later. If what you actually want is an archive rather than editable text, Markdown is the wrong format and this beats everything here, ours included.
Free and already there. Reader view, then export to PDF or rich text. Not Markdown, no files to manage, and for "I want to read this on the plane" it is the whole answer.
Worth doing once
A folder of clipped articles with no source URL and no date is an archive you cannot audit. Six months on you will want to know where a claim came from, and the file will not say. Every tool here can write front matter; set it up on the first clip rather than the hundredth, and include at minimum the source URL and the date you saved it.
PX7 Digital makes PX7 Markdown Web Clipper, so this page is not neutral. Details for the other tools come from their own repositories and sites on 31 July 2026, including the MarkDownload commit date. For archiving rather than reading, SingleFile is the better tool and it is free. Tell us if something is out of date.