☰
Other toolsMarkItDownResearch & Data2026-10-15

Convert HTML to Markdown on a CPU with Microsoft MarkItDown

Our MarkItDown test: on the left, Project Gutenberg ebook #4 HTML (Lincoln's Gettysburg Address); on the right, the Markdown MarkItDown wrote, with the title as a # heading and the body as plain paragraphs
Input: Project Gutenberg ebook #4, public domain in the USA. Output: our MarkItDown run.

We tested Microsoft MarkItDown 0.1.8 on Oct 15: it converted Project Gutenberg ebook #4 (Lincoln's Gettysburg Address, HTML) to Markdown in 0.372 s wall-clock on an 8-vCPU Intel Xeon, CPU only. The title became a Markdown heading and the body became plain paragraphs. MarkItDown is MIT; the ebook is public domain in the USA.

Try it yourself

What you get

Clean Markdown from an HTML page, ready for an AI pipeline or notes app.

Tools

  • Microsoft MarkItDown (MIT)
  • Python package markitdown 0.1.8

Time / cost: 0.372 s wall-clock on an 8-vCPU Intel Xeon, CPU only

Steps

  1. Install it with pip install markitdown (we used markitdown 0.1.8).
  2. Run python -m markitdown input.html -o output.md on your HTML file.
  3. Open the Markdown and check headings, lists and tables by hand; we only tested one Gutenberg HTML page.
  4. For PDF, Word, PowerPoint and other formats, install the extras with pip install 'markitdown[all]'; we only timed HTML → Markdown.

Watch out: MarkItDown is MIT. Input was Project Gutenberg ebook #4 (public domain in the USA). One HTML page only; other formats were not timed.

Proof / source

Tested by AgentPlays Our own test on Oct 15, 2026, on an 8-vCPU machine, CPU only: Project Gutenberg ebook #4 HTML in, Markdown out, and a timing log.

Read the original source ↗ Discussion ↗

Added · How we verify · Report an error

Related