Asva AIPower the future
AEO Glossary · Technical

Canonical URL

The preferred URL for a page, used to avoid duplicate content issues. Important for consistent AI indexing.

What is Canonical URL?

A canonical URL is the single address a site declares as the authoritative version of a page when the same content is reachable at more than one URL. Duplicates arise constantly: http and https, www and non-www, trailing slashes, tracking parameters, print views, paginated variants and syndicated copies. The canonical is declared with a link element in the page head (rel="canonical"), with an HTTP header, or implicitly through redirects and sitemap entries.

Search engines use the declaration as a strong hint about which URL to index and rank and which to fold into it. When canonicals are missing or inconsistent, an engine may index several near-identical pages, split whatever signals they earn between them, and pick a version the site did not intend, such as a parameterised URL or a staging copy.

AI retrieval systems inherit the same problem. An engine that fetches and indexes multiple copies of a page may cite an odd variant, treat the copies as separate weak sources rather than one strong one, or surface stale content from a version the site forgot about. Clean canonicals, backed by redirects from every variant to the preferred URL, mean the engine sees one page with one address and one set of facts.

Why it matters for AI search

AI engines cite URLs, and a citation to a tracking-parameter variant or an old duplicate looks careless and can point users to an out-of-date version. Consolidating duplicates also concentrates retrieval signals on one page, which improves its chance of being chosen over a competitor. Canonical hygiene is unglamorous, but it is part of making sure the engine finds one clear, current answer rather than several conflicting ones.

Related terms

Frequently asked questions

Do AI crawlers respect rel="canonical"?+

Engines that build their own indexes generally treat it as a hint in the same way search engines do, though none document their handling in detail. Redirecting duplicates to the canonical URL is the more reliable approach because it works regardless of how any crawler interprets the tag.

Should the canonical be self-referencing?+

Yes. Every indexable page should declare its own preferred URL as canonical, including the correct protocol, host and trailing-slash form. This prevents parameterised or alternate versions from being treated as separate pages.

More AEO glossary terms

See how AI engines talk about your brand

Track mentions, citations and share of voice across ChatGPT, Perplexity, Gemini and more.

Explore the Brand Visibility Tracker →