Content — Serving Machine-Readable Pages to Agents
The agent-readiness dimension that serves machine-readable, answer-first pages — markdown twins, content negotiation, JSON-LD, meta and canonicals.
What the content dimension means
The content dimension makes the substance of your pages readable by machines, not just rendered for humans. Once an agent has discovered you, it has to extract your actual answer without wrestling layout markup, scripts and navigation chrome. This dimension is about serving the same meaning in a form an agent can lift cleanly — a markdown copy, a negotiated format, structured facts, and unambiguous metadata — and structuring the prose answer-first so the useful sentence is the first sentence.
Signals and standards it covers
- Markdown twins — a
.mdcopy of every HTML page so agents read your content without parsing layout markup. - Content negotiation — returning
text/markdown(or JSON) when an agent sendsAccept: text/markdown, so one URL serves humans and machines the right format. Use the media type registered with IANA. - JSON-LD / schema.org — structured markup so an agent extracts facts as data, not as prose to re-parse.
- Clean meta, canonical and Open Graph — an accurate title and description, a single canonical URL so the agent knows which version is authoritative, and OG tags for consistent representation when your page is referenced.
- Answer-first structure and FAQ schema — leading each section with a standalone declarative answer; answer-first content plus schema is reported as a strong citation signal as of 2026.
How the Agent-Readiness Audit scores it
The Audit scores content on whether an agent can retrieve a clean, structured version of each page. The check content.markdown_twin passes when a page has a reachable .md twin, and content.content_negotiation passes when the same URL returns text/markdown in response to an Accept: text/markdown request. Further checks confirm a single canonical, a present meta description, and at least one valid JSON-LD block. Each criterion is checkable by fetching the URL and inspecting the response — no judgment call. This site proves the dimension live: every page has a .md twin, and requesting this page with Accept: text/markdown returns markdown.
Related: schema.org for agents · serve markdown twins · return text/markdown · get cited by AI (GEO) · audit your site
