How to Save a Web Page as Markdown
Clipping an article into a .md file

What Markdown keeps and what it drops

A clip is a conversion, and conversions lose things. Markdown describes content structure. It has syntax for a heading, a list, a link, an image. It has no syntax for a two-column layout or a sticky sidebar, so those never make it across.

For a long read, an essay, a docs page, or a tutorial, this trade is a good one. The words are what you wanted anyway, and stripping the chrome is half the point. For a pricing table with hover states or a data dashboard, a screenshot serves you better than a clip.

How the clip works, step by step

NoteLoom does this with a Chromium extension. Once it is installed you get an icon in the toolbar, and clipping goes like this:

• Open the article in Chrome, Edge, or Arc. Any normal web page works.
• Optional but useful: open the reading view first, from the toolbar popup or with Alt+R. It runs the text extraction and shows you the article on its own, so you can see whether the right block was caught before you save anything. There is a clip button in the reading view header too, so a good-looking reading view can be clipped straight from there.
• Click Clip to NoteLoom, in the toolbar popup or in that reading view header. The page HTML is converted to Markdown in the browser.
• The clip opens in NoteLoom as a preview. Read it, check the headings survived, and confirm.
• Pick a local folder. The clip is written there as a real .md file, and from that point it is an ordinary text file you can open in any Markdown editor.

The text extraction uses Defuddle, which reads the page structure to find the article body and leave the navigation, ads, and footers behind. Nothing is sent to a server on the way. The extension reads the tab you are on, the conversion runs locally, and the file lands in the folder you authorized.

Where clipping breaks

The clipper only sees what your browser has already rendered, which sets a hard limit on what it can save.

• Login walls and paywalls. If the text is not on the page for you, it is not there for the clipper either. Log in first and the clip works normally.
• Infinite scroll. A feed loads in chunks, so the clip contains the chunks that loaded and nothing below. Scrolling to the end before clipping is the only fix, and on a feed with no end there is no fix.
• Heavy JavaScript pages. If the article is assembled after load, or split across tabs and accordions, the extraction can catch a fragment or grab the wrong block entirely. The reading view tells you this before you save: if it looks wrong there, the clip will be wrong too.
• Pages that are mostly not text. A gallery, a map, an embedded app. The clip technically succeeds and gives you almost nothing.

One more thing worth knowing up front: a clip is the article converted, not condensed. NoteLoom currently does no summarizing or rewriting, so what you save is the full text as the author wrote it, in Markdown. Cutting it down to the two paragraphs you actually cared about is a job for you and the editor. If you wanted a short version, you will be doing that part by hand.

After the file is saved

The clip is a plain .md file in your folder, so everything you normally do with Markdown applies. You can open it in NoteLoom, in VS Code, in any editor at all. Editing it writes straight back to the same file.

A couple of habits that pay off once you have thirty of these: put a source URL and the date at the top of the file, and give the file a name you would recognize in a list. Naming and foldering matter more for clips than for notes you wrote yourself, because in six months you will remember the idea and not the headline.