Why we built it

Almost every long answer from ChatGPT, Claude or Gemini is Markdown underneath: headings, bullet lists, tables, code blocks, sometimes LaTeX. The place that text has to end up is usually a Word document or a PDF, because that is what a professor, a client or a compliance team accepts. Between those two points there was a gap. Copying from the chat window gives you asterisks and pipes; the "export" tools that existed either rendered a screenshot, dropped the tables, or required installing Pandoc and a TeX distribution. We were doing the cleanup by hand several times a week, and we assumed we were not the only ones.

So we built a converter that produces real Word structure: heading styles that show up in the navigation pane, tables you can add rows to, code that keeps its font, and equations that open in Word's equation editor. Then we put it on the pages where people search for it, and in a browser extension on the pages where the answers are.

How it works, in one paragraph

The converter is a small engine written in TypeScript that parses Markdown into a document tree and writes that tree out as OOXML (the format inside a .docx), as paginated PDF, or as HTML. It runs entirely in your browser, which is why there is no upload, no queue and no account for the free features. The same engine ships in the Chrome extension and, for the four cases that need more than a browser tab can give — converting several files at once, merging them into one document, a source file larger than 5 MB, and a PDF you explicitly ask us to render — on our servers, where files live in memory and are deleted within 24 hours. The website you are reading is served by a Go application; the account and billing system runs on Supabase and a payment provider, and nothing else is in the stack.

Who we are

MDtoDOCX is built and run by a small team of engineers and technical writers. Before this we shipped developer tools and document pipelines for larger companies, which is where the habit of checking every generated file in Word, LibreOffice and Google Docs comes from. We are a bootstrapped company: the product is paid for by the people who buy Pro, not by advertising and not by selling data, and we intend to keep it that way.

What we will not do

• We will not read, store or train on the documents you convert. Free conversions never leave your device; Pro server jobs are deleted within 24 hours.
• We will not add countdown timers, fake stock counters or dark-pattern cancellation flows. Prices are on the pricing page and cancelling takes two clicks.
• We will not lock a basic conversion behind a paywall. Headings, tables, code and images stay free.

Contact

Questions, bugs, refund requests and press enquiries all go to [email protected]. We answer within one business day, faster for Pro customers. If a conversion came out wrong, attach the Markdown; that is nearly always enough to reproduce it. Legal documents are at Privacy and Terms.