--- title: "Glossary" description: "Glossary of agent-readability and discoverability terms used on harperreed.com — AGENTS.md, llms.txt, sitemap.md, canonical, JSON-LD, and more." doc_version: "1.0" last_updated: "2026-05-19" canonical: "https://harperreed.com/glossary/" --- # Glossary A short glossary of the terms that show up on this site (and in [`AGENTS.md`](/AGENTS.md)) when describing how AI agents read the web. ## AGENTS.md A plain-text instructions file at the root of a site — like `robots.txt`, but for AI agents and automated systems. Tells crawlers where to find clean content, how to identify themselves, and what to avoid. Lives at [`/AGENTS.md`](/AGENTS.md). ## a14y Short for "agent accessibility." A scorecard and toolkit at [a14y.dev](https://a14y.dev) that measures how well a site is structured for AI agents — separate from WCAG, which measures human accessibility. The name is a homophone, not a synonym. ## canonical link The `` tag in a page's ``, plus the `Link: <…>; rel="canonical"` HTTP header. Both tell agents which URL is the "real" one when the same content is reachable at multiple paths. ## content negotiation When a server returns different versions of the same URL depending on the `Accept` header sent by the client. For example, returning HTML for `Accept: text/html` and Markdown for `Accept: text/markdown` at the same URL. This site does not currently use content negotiation — fetch the [`.md`](/sitemap.md) URL directly. ## JSON-LD JavaScript Object Notation for Linked Data — a way to embed structured data inside an HTML page using a `