Getting Markdown out of a Google Doc

This catches people first. Copying the document's URL, or using "Copy link", gives you an address — there's nothing to convert in that.

What you want is the content on your clipboard. Selecting text in a Doc and copying puts a rich-text HTML version there alongside the plain text, and that HTML carries the headings, lists and tables.

Google's clipboard HTML is not tidy. Almost every element carries a generated class name like c1 or c17, list items get lst-kix_ names, and the whole selection is wrapped in a <b> tag that sets font-weight to normal — a bold tag that turns bold off, which is how the editor tracks document-level formatting internally.

None of that reaches the Markdown. Classes, ids and style attributes are stripped by the sanitiser, because Markdown can't express any of them and clipboard HTML from an arbitrary web page isn't something to trust. The <b> wrapper is unwrapped separately: left in place it would produce a stray ** at the top and bottom of your document.

Links get unwound too. Google routes outbound links through google.com/url?q=… for click tracking inside the editor, and tracking parameters like utm_source often ride along. Both are removed so the link points where it says it points. The output tells you when this happened — changing a link's destination is worth mentioning.

Images in a Doc live on Google's servers. The clipboard HTML references them by URL rather than carrying the bytes, and those URLs are tied to your session — they expire, and they don't resolve for anyone else.

So a pasted document comes through with its text intact and its pictures missing. That's the format's limit, not a setting.

If the images matter, take the other route.

This is the more complete route and it's worth using for anything long. The .docx carries real style information, so headings come from Word styles rather than from a clipboard's best effort, and long documents hold their structure better.

It also means the Google Docs page and the Word page are running the same converter on the same kind of file. Anything true of one is true of the other.

Comments and suggested edits are dropped — you get the document, not its margins. Resolve or accept them before exporting if they matter.

Charts, drawings and smart chips come through as plain text at best. They're objects rendered by the editor, and there's no text equivalent for most of them.

Headers, footers and page numbers go. They belong to a printed page, and Markdown doesn't have pages.

Connecting to Drive would mean asking for access to your files and running a server that holds a token for them. That's a real ongoing risk to you in exchange for saving you one Ctrl+C.

A copy-paste hands over exactly the paragraphs you selected and nothing else. There's no account, no OAuth screen and no add-on to install, and there's nothing on our side that could leak, because there is no our side — the conversion happens in your browser.

Google Docs can also export Markdown itself now. If that output suits you, use it. This exists for the cases where you want to choose the bullet character, the fence style or what happens to images, and for pasting a fragment rather than exporting a whole document.