Opens in a new tabSkip to content
Agent LighthouseAgent Lighthouse

    Searches the text of every published page. The evidence sources themselves are not in this index — search all of them on the trusted sources page.

    GitHub ↗
    Browse checks and page contents
    access-crawl-control/canonical

    Canonical URLs point at the right page

    What it checks

    The canonical URL each page declares, resolved and compared against the page itself — not merely that a tag is present.

    <link rel="canonical"> is read from <head> with rel matched token-wise and case-insensitively, so rel="Canonical", rel=" canonical " and rel="shortlink canonical" all count. Each href is resolved against the page URL (new URL(href, page.url)), so relative canonicals are unambiguous and a malformed value cannot pass as absolute. Comparison runs on one key — host without www., lower-cased path, no trailing slash, query or fragment.

    State Result
    ≥ 2 non-root pages declare the site root as their canonical fail, priority high — homepage collapse
    a canonical does not resolve to an http(s) URL fail, priority high
    most declaring pages collapse onto one non-root URL warn, priority high
    a page carries two conflicting canonicals warn, priority medium
    a canonical names another domain warn, priority medium
    some pages do not declare one, and every declared one is sound pass, naming the pages without one
    every page that declares one resolves to itself pass
    no page declares a canonical in <head> na — nothing to judge
    no pages scanned na

    Why it matters

    Googlebot reads <link rel="canonical">, and uses it as a strong signal when choosing which of a set of duplicate URLs to index and consolidate signals onto. The URL it selects is the one eligible to be shown in Search — and so, in turn, as a supporting link in AI Overviews and AI Mode.

    Evidence

    • Google describes rel="canonical" as “a strong signal that the specified URL should become canonical”. It explains the effect: “It helps search engines to be able to consolidate the signals they have for the individual URLs (such as links to them) into a single, preferred URL.” It also recommends a self-referential canonical on the canonical page — developers.google.com/…/consolidate-duplicate-urls (verified 2026-08-21)
    • RFC 6596, “The Canonical Link Relation” (Informational, April 2012), defines the relation type. It designates “an Internationalized Resource Identifier (IRI) as preferred over resources with duplicative content”. That is a published specification behind the attribute, with known consumers — rfc-editor.org/…/rfc6596.html (verified 2026-08-21)
    • Google’s crawl-budget guidance corroborates the crawling-side effect: “Consolidate duplicate content” to “focus crawling on unique content rather than unique URLs”, because otherwise duplicate URLs waste “a lot of Google crawling time on your site” — developers.google.com/…/large-site-managing-crawl-b… (verified 2026-08-21)
    • The link from indexing to AI surfaces is documented: “To be eligible to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Google Search with a snippet” — developers.google.com/…/ai-features (verified 2026-08-21)

    Limits

    Google explicitly states the tag is not required: “none of them are required; your site will likely do just fine without specifying a canonical preference” (developers.google.com/…/consolidate-duplicate-urls, verified 2026-10-17). The signal is a hint, not a directive, and Google may select a different canonical than the one declared. The same page says why: “That’s because if you don’t specify a canonical URL, Google will identify which version of the URL is objectively the best version to show to users in Search.” It also calls the tag “generally not critical” (developers.google.com/…/consolidate-duplicate-urls, verified 2026-10-17).

    Google’s canonicalization overview describes the fallback: Google “chooses the page that, based on the factors (or signals) the indexing process collected, is objectively the most complete and useful for search users, and marks it as the canonical” (developers.google.com/…/canonicalization, verified 2026-10-17). No primary source says a missing canonical costs a site anything. Absence of the tag is therefore not a defect, and the audit does not score it. When no scanned page declares a canonical, the audit is na: it grades the canonical’s value, and there is no value to grade.

    When some pages declare one and others do not, only the declared ones are judged, and the pass message names the pages without one. The graded mechanism concerns the canonical’s value, and that is where the audit fails. Two or more pages that declare the site root as their canonical fail as a homepage collapse, and a canonical that does not resolve to an http(s) URL fails. Most declaring pages collapsing onto one other URL warns at high priority.

    RFC 6596 is Informational, not Standards Track (rfc-editor.org/…/rfc6596.html, verified 2026-08-21). No AI vendor outside Google names the signal. OpenAI’s (developers.openai.com/…/bots) and Perplexity’s (docs.perplexity.ai/…/perplexity-crawlers) crawler documentation mention robots.txt only (both verified 2026-08-21), and Google states no special markup is needed for its generative features (developers.google.com/…/ai-optimization-guide, verified 2026-08-21). The audit only reads <link rel="canonical"> in <head>. A canonical sent only in a Link: response header is not read, so that site is na, as if it declared none.

    How it scores

    A named crawler’s use of the signal is stated in vendor documentation, and the link relation itself is a registered, published standard.

    Sources