HTML to Markdown Converter with GFM Tables and Code Blocks

Convert HTML to Markdown for docs, READMEs and CMS migrations

Paste HTML on the left and get Markdown on the right, updated as you type. Typical uses are moving blog posts out of WordPress or another CMS into a static site generator, turning a rendered web page or email into notes, pasting documentation into a GitHub README or wiki, and cleaning up HTML that a rich-text editor produced so it can live in a Git repository as plain text.

The converter parses your input with the browser's DOMParser, the same HTML parser that renders web pages, so messy real-world markup — unclosed <p> and <li> tags, uppercase tag names, attributes without quotes — is understood the way a browser understands it. Parsing in this mode does not run scripts or load images, and <script>, <style>, <noscript>, <iframe> and form controls are dropped from the output entirely. Nothing is uploaded; the HTML to md conversion happens on your device.

What gets converted, and how

Headings h1 to h6 become # to ###### (ATX style), or, if you prefer, h1 and h2 can be written in Setext style with === and --- underlines. Paragraphs are separated by blank lines and <br> becomes a hard line break. <strong> and <b> become **bold**, <em> and <i> become *italic* (or _italic_), and <del> or <s> become ~~strikethrough~~.

Inline <code> becomes backtick code, using a longer run of backticks when the code itself contains one. <pre> blocks become fenced code blocks, and the language is kept when the markup carries it as class="language-js", class="lang-js" or a data-language attribute — the conventions used by Prism, highlight.js and most static site generators — so syntax highlighting survives the trip. Code inside <pre> is copied exactly as written, with its indentation and blank lines intact, and is never escaped, since Markdown does not interpret anything inside a fence. Links become [text](url "title") and images ; URLs containing spaces or parentheses are wrapped in angle brackets so they don't break.

Unordered and ordered lists are converted with correct nesting and indentation, ordered lists respect a start attribute, and you can choose -, * or + as the bullet. Blockquotes become > lines, including nested content. Tables become GitHub-flavored Markdown tables, with the first row as the header, column alignment taken from align or text-align, and pipes inside cells escaped. <hr> becomes ---. Characters that Markdown would otherwise treat as formatting, such as * and _ in ordinary text, are backslash-escaped so the rendered result reads the same as the original.

Limits worth knowing

Markdown is a smaller language than HTML, so some things cannot survive. Inline styles, classes, colors and layout (columns, floats) are discarded. Tables with merged cells (colspan or rowspan), or cells containing lists and multiple paragraphs, are flattened to one line per cell because GFM tables cannot hold block content. Elements with no Markdown equivalent — <span>, <div>, <section> — are unwrapped and their contents kept.

To go the other way, use the Markdown to HTML converter, and use the Markdown preview tool to see how the converted Markdown renders before you commit it.