The short version
An HTML to Markdown converter maps HTML elements onto Markdown equivalents: headings become hashes, strong and em become asterisks, anchors become bracket-paren links, lists become dashes or numbers, and tables become pipe tables. Presentational CSS has no Markdown equivalent and is discarded.
Why this one runs in your browser
Because pasted HTML frequently comes from somewhere you would rather not upload. An internal wiki, a client's site, an email template, a page behind a login.
There is no reason for that to touch a server. Converting HTML to Markdown is pure text transformation, so it runs in the browser tab and the content never leaves your machine. We log that a conversion happened, never what was converted.
The one thing that does need a server is fetching a live URL, which is why our webpage-to-Markdown tool is separate.
The mapping
Most of HTML has an obvious Markdown counterpart, and a handful of things do not.
| HTML | Markdown | Notes |
|---|---|---|
| h1 to h6 | # to ###### | Level preserved |
| strong, b | **bold** | |
| em, i | *italic* | |
| a href | [text](url) | Empty links are dropped |
| img |  | Alt text preserved |
| ul, ol | - item, 1. item | Ordered lists get real numbers |
| table | Pipe table | First row becomes the header |
| pre, code | Fenced or inline code | Entities decoded |
| blockquote | > quote | |
| div, span, CSS | Discarded | No Markdown equivalent |
Where conversion is genuinely lossy
Any meaning carried by presentation is lost, because Markdown has no way to express it. A two-column layout becomes one column. Colour-coded text becomes plain text. A carefully spaced pricing table becomes a pipe table.
Deeply nested structures flatten. HTML can nest divs arbitrarily; Markdown's structure is much shallower, so a complex layout becomes a linear document.
Custom attributes, data attributes, ARIA roles and inline event handlers all disappear. If you are converting HTML that does something rather than just says something, Markdown is the wrong destination.
Cleaning up afterwards
Converted Markdown from a real web page usually needs a pass by hand, and knowing what to look for saves time.
Check heading levels first. Sites frequently use an h3 for visual size rather than document structure, so the output can start at the wrong level or skip levels.
Then check links. Relative hrefs stay relative, which will be broken wherever you paste the result. Converting them to absolute URLs is usually what you want.
Finally look for stray empty list items or table rows, which come from markup that had a structural element with no content in it.
Questions about this tool
Is my HTML uploaded anywhere?+
No. This tool runs entirely in your browser tab. We record that a conversion happened, never the content.
Does it handle tables?+
Yes, as GitHub-flavoured pipe tables, with the first row treated as the header. Complex tables with merged cells will flatten, since Markdown has no equivalent.
What happens to CSS and styling?+
It is discarded, because Markdown describes structure rather than presentation. Anything whose meaning depends on layout or colour will be lost.
Why are my links broken after converting?+
Relative hrefs stay relative, so they break wherever you paste the output. Convert them to absolute URLs, which is usually what you want anyway.