HTML to Markdown

Adds Paste as Markdown to text you copy with its formatting, from a web page, Google Docs, Notion, Word or Outlook. The row shows the Markdown before you pick it, and that Markdown is what gets pasted: headings, lists, links, tables and code come out as clean Markdown for your notes, READMEs and docs.

Example

Copy this from a web page:

Release notes

• Faster search, see the guide
• Fixed a crash in sync

Press ⇧ on the item and choose Paste as Markdown:

**Release notes**

- Faster search, see [the guide](https://example.com/guide)
- Fixed a crash in `sync`

What it converts

• Headings, paragraphs and line breaks.
• Bold, italic, strikethrough and inline code, from tags and from the inline styles editors use (font-weight: 700, font-style: italic, text-decoration: line-through).
• Links with their titles, and images with their description, address and title. A link whose text is its own address becomes <https://…>.
• Bulleted, numbered and nested lists, with each list's starting number, and task lists (- [x]).
• Block quotes, horizontal rules and code blocks, with the language when the page names it (language-js, lang-js, highlight-source-js, data-lang).
• Tables, with the column alignment the page sets. Cells that span columns or rows leave the cells they cover empty.
• Entities such as & and ’, and white space the way a browser shows it.
• Characters Markdown could read as syntax get a backslash: *, `, [ and ] always, and the start of a line that would read as a heading, a list item, a quote or a table rule (#, -, 1., >, -- | --). An underscore inside a word, as in snake_case, stays as it is.

It knows the quirks of the apps you copy from:

• Google Docs: the whole copy is wrapped in bold that isn't bold, nested lists sit beside their parent item, and checklists keep their ticks.
• Notion: to-do lists keep their state, and lists split into one item each are joined again.
• Word and Outlook: bullets and numbers written as paragraphs become real nested lists, and the stylesheet, conditional comments, <o:p> and hidden text are left out.
• Safari and Chrome: bold and italic written as inline styles count, and the styles a browser adds to everything do not.
• Code editors: code copied as colored lines in a monospaced font becomes a code block.

Limits

• It needs the copy's formatting. Plain text, and copies whose formatting is larger than 256 KB (yabbistir doesn't pass those to extensions), get the message This copy has no formatting to turn into Markdown.
• Relative links stay as they are, since the page they came from isn't known.
• Images stored inside the copy, rather than at an address, are left out; their description stays as text. Images copied from Word point to temporary files on your Mac.
• Underline, highlight, colors, font sizes, and superscript and subscript have no Markdown, so that text stays plain.
• A table cell is one line: its paragraphs and line breaks become <br>. A table used for layout, such as an email's columns or a cell holding a list or heading, is read as its content instead.
• In Google Docs and Word, text in a monospaced font such as Courier is code: a paragraph of it becomes a code block, a few words become inline code. When the whole copy is in that font, it stays text. On web pages, only real code elements count.
• Markdown cannot show a few combinations, such as bold that ends in punctuation right before a letter (**Note:**then); those may lose their bold.

Permissions

None. The extension reads only the item you pick, and converts it on your Mac.

Data

The names of HTML's character entities come from the WHATWG HTML Standard, used under CC BY 4.0.