How do I convert a Confluence page to Markdown?

Open the page, select the body, copy, then paste into the box below. The clipboard carries the page as HTML and that is what gets converted, so headings, tables, lists and links come across as Markdown rather than as flattened text.

Does it handle Confluence tables?

Yes, as long as they are ordinary tables. They become GitHub-flavoured Markdown tables with pipes. Merged cells are the exception: Markdown has no way to express a cell spanning two columns, so those come out as separate cells and need a manual tidy.

What happens to macros, panels and status lozenges?

They have no Markdown equivalent, so they degrade to their text content. Info and warning panels usually become plain paragraphs or blockquotes; a table of contents macro produces nothing, because the content was never in the page: Confluence generated it at render time.

Can I export a whole space at once?

Not from here. This converts one page at a time from your clipboard. For a bulk migration, Confluence's own space export produces HTML or XML, and the HTML files can then be converted one by one on the HTML to Markdown page.

Can I write Markdown inside Confluence?

Partly. The cloud editor converts Markdown as you type (## then a space becomes a heading, ** ** becomes bold) but it stores rich text, not Markdown, so the file is never Markdown afterwards. There is also an Insert > Markdown option that takes a block of it in one go, and the older Data Center wiki markup is a different syntax again.

What about attachments and images?

Image references come across pointing at their original Confluence URLs, which usually require a login to load. Download the ones you need and repoint the links, or delete them: the Markdown itself is unaffected either way.

Confluence to Markdown

Copy a Confluence page, paste it below, and get Markdown back: headings, tables, lists and links intact. No API token, no admin rights, and nothing leaves your browser.

Paste your Confluence page below, or drop an exported .html file

Select the page body, copy, then paste

Paste or type here

Markdown

Converted in this tab. The file is never uploaded.

Why copy and paste rather than an export

It is the shortest route that actually preserves structure. When you copy from Confluence, the clipboard holds an HTML version of the selection alongside the plain text: the same HTML the browser was rendering. This page converts that, so a heading stays a heading instead of becoming a line of text that merely looks important.

The alternative, Confluence's space export, is the right tool for moving hundreds of pages at once. It produces HTML or XML files, which you can convert here through HTML to Markdown. For one page, or ten, the clipboard is faster and needs no permissions.

What survives, and what cannot

Survives: headings, paragraphs, bold and italic, ordered and unordered lists including nesting, links, inline code, code blocks, blockquotes and ordinary tables.
Degrades to text: panels, info and warning boxes, status lozenges, expand blocks. Markdown has no equivalent, so the words inside them survive and the decoration does not.
Disappears: macros that generate content at render time, such as a table of contents or a page-tree listing. There was nothing in the page to copy; Confluence assembled it on the way out.
Needs attention: merged table cells, which Markdown cannot express, and image links, which still point at Confluence.

Writing Markdown in Confluence, the other direction

The question people ask alongside this one is whether Confluence understands Markdown at all. It does, in three limited ways, and none of them mean a Confluence page is ever stored as Markdown.

Autoformatting as you type. In the cloud editor, typing ## makes a heading, **bold** makes bold, - starts a list and ``` opens a code block. The characters are consumed and replaced with rich text: what you get is a formatted page, not a Markdown one.
Insert → Markdown. Paste a whole block of Markdown into one dialogue and Confluence converts it in place. This is the right route for a README or a set of release notes, and it is a one-way trip.
Wiki markup, on Data Center and Server. An older syntax of Confluence's own: h1. for a heading, *bold* with one asterisk: that predates Markdown and is not compatible with it.

So Markdown is an input format in Confluence and never a storage format. That is precisely why getting a page back out as Markdown needs the conversion above: there is no original Markdown to recover, only rendered HTML to translate.

Then tidy it

Confluence's HTML tends to be verbose, so the converted Markdown often has inconsistent bullets and stray blank lines. The formatter normalizes all of that without changing what the document says, and the editor shows you the rendered result while you fix anything the conversion could not.

Common questions

How do I convert a Confluence page to Markdown?: Open the page, select the body, copy, then paste into the box below. The clipboard carries the page as HTML and that is what gets converted, so headings, tables, lists and links come across as Markdown rather than as flattened text.
Does it handle Confluence tables?: Yes, as long as they are ordinary tables. They become GitHub-flavoured Markdown tables with pipes. Merged cells are the exception: Markdown has no way to express a cell spanning two columns, so those come out as separate cells and need a manual tidy.
What happens to macros, panels and status lozenges?: They have no Markdown equivalent, so they degrade to their text content. Info and warning panels usually become plain paragraphs or blockquotes; a table of contents macro produces nothing, because the content was never in the page: Confluence generated it at render time.
Can I export a whole space at once?: Not from here. This converts one page at a time from your clipboard. For a bulk migration, Confluence's own space export produces HTML or XML, and the HTML files can then be converted one by one on the HTML to Markdown page.
Can I write Markdown inside Confluence?: Partly. The cloud editor converts Markdown as you type (## then a space becomes a heading, ** ** becomes bold) but it stores rich text, not Markdown, so the file is never Markdown afterwards. There is also an Insert > Markdown option that takes a block of it in one go, and the older Data Center wiki markup is a different syntax again.
What about attachments and images?: Image references come across pointing at their original Confluence URLs, which usually require a login to load. Download the ones you need and repoint the links, or delete them: the Markdown itself is unaffected either way.

Sources

Atlassian: Confluence wiki markupthe storage format the export leaves behind