How to save a web page as Markdown

Last updated 13 October 2026

Converting a web page to Markdown sounds like it should be a single button. In practice there are three ways to do it, and they fail in different places.

The short answer

One-off, simple article, no formatting to preserve → copy from the page and paste; most editors convert it acceptably.
Occasional, and you don't mind cleaning up → a free converter extension. Expect to delete the navigation and footer by hand.
Regular reading habits, or the page has tables and code → a dedicated clipper. This is the only category that reliably keeps structure intact.

The deciding factor isn't the converter — it's what the page contains. A plain blog post converts fine with almost anything. A documentation page with a comparison table and a code sample is where most tools fall apart.

Method 1: copy and paste

You select the article, copy it, and paste into your editor. Some editors (Obsidian, Typora, VS Code with a Markdown extension) convert the clipboard's rich text automatically. It is genuinely fine for short, simple pages.

Where it breaks:

You select by hand — and you'll usually miss the last paragraph or grab the "related posts" block.
Code blocks often arrive as inline text, losing indentation and the language tag, so syntax highlighting is gone.
Links usually survive. Images may or may not, depending on the editor.
Tables are the worst case: a real table becomes a run of lines separated by spaces.

Method 2: a generic converter extension

These take the full HTML of the page and convert it. Many are free and open-source. They're reliable, they work offline, and they never touch your account.

Where it breaks: they convert the whole page. That means you get the navigation bar, the cookie notice, the newsletter prompt, the ad slots and the copyright footer in your note, and you delete them by hand afterwards. On pages with heavy chrome — most news sites — you spend longer cleaning than reading.

Some of these tools let you pass a CSS selector to target the article. That works well, but it means writing a rule for every site you read.

Method 3: a clipper that extracts the article first

This category runs readability-style extraction over the page before converting, so only the article body reaches the converter. The navigation, ads and popups are removed at the source, not by you afterwards.

That's the approach Clipperry takes. It saves to Notion, Obsidian or a plain .md file, and the parts that usually break are the parts it handles explicitly:

Tables come out as real Markdown tables with the pipes intact, including empty cells.
Code blocks keep their language identifier, so python stays python and highlighting survives.
Images are rewritten to absolute URLs so they don't break when you open the note months later.
Metadata — title, source URL, date and tags — is filled in automatically.

What a clip looks like

---
title: "How to save a web page as Markdown"
source: "https://example.com/article"
date: 2026-10-13
---

## The article's own heading

Body text, with **emphasis** and [links](https://example.com) preserved.

| Option            | Keeps tables | Keeps code language |
|-------------------|--------------|---------------------|
| Copy-paste        | No           | No                  |
| Generic converter | Sometimes    | Rarely              |
| Article clipper   | Yes          | Yes                 |

What to check before you pick one

Test on your worst page. Not a blog post — find the page you clip most often. If it has a table, use that one. Every tool looks good on a simple page.
Check where the file goes. Some tools only export a download; others write into your vault or database. If you clip while researching, a download folder becomes a graveyard.
See whether it remembers your destination. Re-selecting a Notion database every single time is why people quietly stop using a clipper after a week.
Read what it sends over the network. Local extraction means the page never leaves your browser. Server-side extraction means it does — which is a reasonable trade for some people, but you should know which one you're choosing.

Frequently asked questions

Can I convert a saved HTML file instead of a live page?

Most converter extensions work on the currently open tab, so opening the local file in Chrome and running the clipper on it usually works. Tools that call an API need a public URL and won't accept a local file.

Why do my tables turn into plain text?

Usually one of two reasons: the tool converted the rendered text rather than the table element, or it dropped empty cells so the columns stopped lining up. Markdown tables need the pipes in every row, even when a cell is blank.

Does the code block keep its language?

Only if the tool reads the class on the code element. Tools that work from rendered text lose it, which is why pasted snippets come back with no syntax highlighting.

Is there a way to do this without an extension?

Yes. Developer APIs such as Jina Reader convert a URL to Markdown over HTTP, and pandoc handles local HTML files. Both are good options if you're scripting rather than reading.

Try it

Clipperry is a Chrome extension. The free plan covers 50 clips a month and includes every destination, so you can test it on the page you actually care about before deciding anything.

Get Clipperry free →