HTML to Markdown

Paste HTML on the left and Markdown appears on the right. Swap the boxes to go back the other way.

Script and style elements are dropped rather than converted, because their contents are code that would otherwise land in your document as text. Markdown's own markers are escaped in the output, so an asterisk in the HTML comes out as \* and doesn't turn into emphasis on the way back.

HTML can say more than Markdown can. A span with a class, a nested table, a definition list, an inline style, an id used as an anchor target: none of these has a Markdown spelling, so the converter keeps the text and drops the wrapper.

Whitespace goes too. HTML collapses runs of spaces when it renders, so this does the same, and a document that relied on preformatted spacing outside a pre element won't come back the same. Text inside a pre element keeps its spacing and blank lines.

Markdown was designed around that trade. Gruber's stated goal was that the source should read as plain text, so he left out whatever can't be written plainly.

Markdown came out of a collaboration between John Gruber and Aaron Swartz in 2004. Swartz, then 17, worked on the syntax with Gruber and wrote html2text, a Python script that did exactly what this page does.

It turned a web page into plain text that was also valid Markdown. His script lives on as the html2text package on PyPI, which still credits him and last released in April 2025.

Swartz also wrote atx, the heading format Markdown borrowed, and worked on RSS 1.0 and the early Creative Commons licences. He died in 2013, aged 26, while facing federal charges over bulk-downloading academic papers from JSTOR.

Markdown to HTML is the forward direction, and it explains why the two formats disagree about so much. XML to JSON uses the same tag parser underneath, pointed at data rather than at prose.

Text rescued from a web page often needs line breaks removed before it reads properly.