Web Content & Formats

What is Markdown?

A lightweight plain-text formatting syntax that converts deterministically to HTML, created in 2004 to be readable as source and as rendered output.

Markdown uses punctuation that already looks like formatting in plain text: # heading, **bold**, _italic_, - list, > quote, ` code `. A renderer parses those conventions and emits the equivalent HTML. The result is a format you can read in a terminal, write in any text editor, and render anywhere.

For LLMs and AI pipelines specifically, Markdown is the preferred document format. It's denser than HTML (no tag overhead), preserves structure (headings, lists, tables) that vanilla text loses, and is what most models were trained on. Most modern scraping APIs default to returning Markdown rather than HTML for that reason.

GitHub Flavored Markdown (GFM) added tables, task lists, autolinks, and fenced code blocks; CommonMark stabilized the spec. Almost every Markdown parser in 2026 supports both.

In the wild

  • # Title<h1>Title</h1>
  • [Brand.dev](https://brand.dev) → an anchor
  • js\nconst x = 1;\n `, a fenced code block

How Brand.dev uses markdown

Endpoints in the Brand.dev API where this concept comes up directly.

FAQ

Markdown vs HTML?

Markdown is shorter, more readable as source, and renders to HTML. HTML is more powerful (any element, any attribute) but more verbose. Use Markdown for content, HTML for layout.

Why do AI tools prefer Markdown?

It's structural without being noisy, denser than HTML, and matches the format models saw most often during training. Cleaner Markdown in → cleaner answers out.

Is there a Markdown standard?

CommonMark is the closest thing to a standard. GitHub Flavored Markdown extends it with tables and tasks; almost every renderer today supports both.

Related terms

Ship an agent that actually knows things.

Free tier, 10-minute integration, and the same API powering agents at Mintlify, daily.dev, and Propane. No credit card to start.