To strip HTML tags that don't have a Markdown equivalent while preserving the content inside them, set strip_tags to true, like this:

$converter = new HtmlConverter(array('strip_tags' => true));

$html = 'Turnips!';

$markdown = $converter->convert($html); // $markdown now contains "Turnips!"

Or more explicitly, like this:

$converter = new HtmlConverter();

$converter->getConfig()->setOption('strip_tags', true);

$html = 'Turnips!';

$markdown = $converter->convert($html); // $markdown now contains "Turnips!"

Note that only the tags themselves are stripped, not the content they hold.

To strip tags and their content, pass a space-separated list of tags in remove_nodes, like this:

$converter = new HtmlConverter(array('remove_nodes' => 'span div'));

$html = 'Turnips!

';

$markdown = $converter->convert($html); // $markdown now contains ""

Style options

Bold and italic tags are converted using the asterisk syntax by default. Change this to the underlined syntax using the bold_style and italic_style options.

$converter = new HtmlConverter();

$converter->getConfig()->setOption('italic_style', '_');

$converter->getConfig()->setOption('bold_style', '__');

$html = 'Italic and a bold';

$markdown = $converter->convert($html); // $markdown now contains "_Italic_ and a __bold__"

Limitations

Markdown Extra, MultiMarkdown and other variants aren't supported – just Markdown.

Known issues

Nested lists and lists containing multiple paragraphs aren't converted correctly.

Lists inside blockquotes aren't converted correctly.

Report your issue or request a feature here. Issues with patches or failing tests are especially welcome.

Style notes

Setext (underlined) headers are the default for H1 and H2. If you prefer the ATX style for H1 and H2 (# Header 1 and ## Header 2), set header_style to 'atx' in the options array when you instantiate the object:

$converter = new HtmlConverter(array('header_style'=>'atx'));

Headers of H3 priority and lower always use atx style.

Links and images are referenced inline. Footnote references (where image src and anchor href attributes are listed in the footnotes) are not used.

Blockquotes aren't line wrapped – it makes the converted Markdown easier to edit.

Dependencies

HTML To Markdown requires PHP's xml, lib-xml, and dom extensions, all of which are enabled by default on most distributions.

Errors such as "Fatal error: Class 'DOMDocument' not found" on distributions such as CentOS that disable PHP's xml extension can be resolved by installing php-xml.

Contributors

Many thanks to all contributors so far. Further improvements and feature suggestions are very welcome.

How it works

HTML To Markdown creates a DOMDocument from the supplied HTML, walks through the tree, and converts each node to a text node containing the equivalent markdown, starting from the most deeply nested node and working inwards towards the root node.

To-do

Support for nested lists and lists inside blockquotes.

Offer an option to preserve tags as HTML if they contain attributes that can't be represented with Markdown (e.g. style).

Trying to convert Markdown to HTML?

Use one of these great libraries:

No guarantees about the Elvish, though.