PDF to Markdown, Word to Markdown, and Six More Converters That Run in Your Browser

Most "free converter" sites work the same way: you upload your file to a server you know nothing about, wait, and download a result. For a marketing PDF that is fine. For a contract, an internal spec, or a document with customer names in it, it is a data-handling decision nobody asked you to make.

The converters on dochost run in your browser. The file does not leave your machine, there is no upload step, and there is no account.

What is available

Into Markdown

• PDF to Markdown — keeps headings, lists and tables where the PDF has a real text layer
• Word to Markdown — .docx to clean Markdown
• HTML to Markdown and URL to Markdown
• CSV to Markdown and HTML table to Markdown
• Confluence to Markdown

Out of Markdown

• Markdown to PDF
• Markdown to Word and Markdown to Excel
• Markdown to Google Docs and Markdown to Notion
• Markdown to rich text, for pasting into an editor that does not speak Markdown

The full set lives on the tools hub.

The scanned-PDF caveat, stated up front

A PDF that is only scanned images has no text layer, so there is genuinely nothing to convert. Rather than return an empty file and let you think the tool is broken, PDF to Markdown tells you that is what happened. A partly image-only PDF converts what it can and reports how much it skipped.

We would rather say "this page had no text" than hand back a silently empty document.

Convert, then share

Every converter output can go straight into a page: convert a .docx to Markdown, check the preview, and publish it as a link in the same session. That is the whole reason the converters live here rather than as a separate site — the common job is not "convert a file", it is "get this document in front of someone".

If you only need the file, take the download and close the tab. Nothing is stored.

Why local conversion matters more than it used to

Documents increasingly pass through a model before they reach a person. That means more conversions, on more sensitive material, done casually. A converter that runs locally removes the question of what a third party retained, because there was no transfer to retain.

It also means the tools work offline once the page has loaded.

What survives each conversion, and what doesn't

Converting between formats always loses something. The useful thing is knowing what before you rely on the output. These are the limits each converter reports, rather than leaving you to discover them:

Two rules of thumb fall out of that table. Headings survive when the source used real heading styles, and not when someone just made the text bold and big. And images never travel inside Markdown, because Markdown can only point to an image at a URL.

One exception to "nothing leaves your machine"

URL to Markdown is different from the rest, and it should say so. To read a web page, dochost has to fetch it, and that fetch runs on our server so the browser's cross-site restrictions don't block it. The conversion to Markdown still happens in your tab. The page you point it at is public by definition, so nothing private crosses, but "offline once loaded" does not apply to this one.

Convert, check, publish: the round trip

The usual reason to convert is to get a document in front of someone who will read it in a browser. The recording below is the last step: a Markdown file is uploaded and comes back as a page with its own link.

Recorded on dochost.io: a .md file uploaded and published as a formatted page.

If the source was a .docx or a PDF, the steps before this one are the converter above, then a skim of the preview to catch a table that flattened or a heading that came through as a bold paragraph.