Конвертируйте HTML в Markdown
Tool guide
You exported an article from a CMS, or pulled a block out of an old email template, and now it has to live in a repository where everything is Markdown. Paste the markup into the HTML Input box and press the convert button: headings turn into hash marks, list items into dashes, anchors into bracketed links. The parsing runs in your own tab, so nothing about the pasted markup leaves the browser. See also: render that Markdown back to HTML, save the finished Markdown as a PDF, strip the page down to plain text.
Don't count on tables. A table element with its rows and cells is a much richer structure than a line of pipe characters, and the output nearly always needs hand-editing. If your source is table-heavy, rebuild them separately, or simply leave the original HTML inside the Markdown file — the spec allows it and GitHub renders it.
They vanish along with the tags. Markdown describes the shape of text, not its appearance, so there is nowhere to put a class name or a CSS property. What survives is whatever the syntax can express: heading level, list item, bold, italic, link target. Keep the original HTML alongside the .md if the styling matters.
The download button writes the output box to a plain text file in UTF-8, so non-Latin characters come through intact. The filename is fixed by the page rather than chosen by you, so rename it for your project after it lands in your downloads folder — README.md, CHANGELOG.md, whatever fits.
An img tag becomes an image link with a leading exclamation mark, but no file is copied anywhere — only the address travels. If the pictures sat on the site under relative paths like /upload/img.png they will not resolve inside a repository, so swap those for absolute URLs or place the files next to the document.
There is no file picker on the page; pasting into the HTML Input box is the only way in. Open the file in an editor, copy the contents, paste them here. For a live page, view-source works, though it pays to grab only the block you need — headers, menus and footers otherwise arrive as noise in your Markdown.
It goes nowhere. The conversion is done by a script on the page itself, and no network request carries your text. Nothing is stored in a database or written to a server file, and the boxes empty on reload. That is also why it keeps working offline once the tab has loaded.
No hard limit is coded in, but the parsing happens on the tab's single thread. An article fragment converts instantly; a saved page of several megabytes will freeze the interface for a second or two. Large exports are easier to handle section by section, and the result is easier to proofread that way too.
Before: InstallClone the repository.
After: ## Install Clone the [repository](https://example.com/repo).
Before: Node 18Git
After: - Node 18 - **Git**
Your rating and feedback help decide what to improve next.
