Comparison
Schema.org vs Markdown for Agents: facts or content
Schema.org describes structured facts; Markdown serves clean content for LLMs. Different layers — they work together.
In short
Not a choice but two layers: Schema.org gives a machine structured facts (type, price, rating), Markdown for Agents gives clean content with no HTML noise. Together an LLM gets both structure and readable text.
What each is
- Schema.org JSON-LD — entity markup in a
<script type="application/ld+json">. A machine understands: this is aProduct, the price is X. - Markdown for Agents — serving a Markdown
version of a page on
Accept: text/markdown. Dense, clean text instead of HTML.
Side by side
| Schema.org | Markdown for Agents | |
|---|---|---|
| Describes | Facts / entities | Content / prose |
| Format | JSON-LD in <head>/<body> | a text/markdown body |
| How consumed | Parsing markup | Reading text |
| Strong for | Rich results, facts | Understanding & citing content |
Key difference
Schema.org is about facts to parse (price, author, date). Markdown is about content to read (the whole article). The first structures; the second eases reading.
Do both
Yes — they’re at different levels: Schema.org puts facts in markup, Markdown serves the clean body. An AI engine gets both precise attributes and citation-friendly text.
How to verify
A scan checks for both Schema.org JSON-LD and Markdown content negotiation — both signals in one report.