A canonical URL is the version of a page declared as the authoritative one when multiple URLs serve duplicate or near-duplicate content. It is set with a link tag in the page head: <link rel="canonical" href="https://example.com/canonical-version/">.
Canonical URLs exist because URL-level duplication is a normal consequence of the web — tracking parameters, session IDs, www vs non-www, http vs https, trailing slash differences, faceted navigation, paginated lists. Without canonicalization, search engines must guess which version to index and may split ranking signals across multiple URLs, weakening all of them.
For SEO, the rules are: every important page should have a self-referencing canonical (a page pointing at itself), duplicate variants should canonicalize to the primary URL, and the canonical should always be absolute, https, and match the URL Google would actually crawl. Mismatched canonicals (declaring a different page than the one Googlebot fetched) are silently ignored.
For AEO, canonical URLs are equally important — AI engines deduplicate by canonical to avoid citing the same content under multiple URLs. A site with broken canonicalization may be cited under the wrong URL or split across multiple variants, fragmenting its AI visibility. Pair canonical declarations with consistent internal linking to reinforce the canonical for crawl bots.