Canonical tag: definition and how it works

99Visibility Research · Updated October 21, 2026 · 4 min read

Short answer

A canonical tag is a link element that tells search engines which URL is the preferred version of a page when several URLs show the same or similar content. It is written as rel="canonical" in the page head or in an HTTP header. Engines treat it as a signal, not a command. The URL they pick as canonical is the one they index, and so the one that can appear as a source.

Key takeaways

A canonical tag names the preferred URL among duplicates, using rel="canonical" in the HTML head or an HTTP Link header.
It is a signal that search engines weigh against other evidence, so the URL they choose can differ from the one you declared.
Google requires a page to be indexed and snippet-eligible to appear as a supporting link in AI Overviews or AI Mode, so a canonical conflict can decide which URL is eligible.
A canonical tag is not a ranking or citation lever. It decides which duplicate counts, not whether any of them gets cited.

Definition

A canonical tag declares the canonical URL: the single address you want search engines to treat as the main copy of a piece of content. Duplicates happen for ordinary reasons. Tracking parameters, sort orders, print versions, http and https variants, and trailing slashes all create extra URLs for one page.

The tag sits in the page head:

<link rel="canonical" href="https://example.com/pricing">

For files without an HTML head, such as PDFs, the same signal can be sent as an HTTP header:

Link: <https://example.com/guide.pdf>; rel="canonical"

This entry is a short definition. For the failure modes and how to diagnose them, read whether canonical tag mistakes can keep pages out of AI Overviews.

How it works

A crawler fetches a URL, reads the declared canonical, and compares it with other evidence. That evidence includes internal links, redirects, sitemaps and how similar the pages actually are. The engine then groups the duplicates and picks one URL to index and show.

When your signals agree, the engine usually picks the URL you declared. When they disagree, it may pick another one. Common conflicts:

The canonical points to a URL that redirects, returns an error, or is blocked.
Internal links and the sitemap point to a different variant than the canonical.
Two pages each declare the other as canonical.
A page declares itself canonical and also carries noindex.

The fix is consistency. Pick one URL per piece of content, then make the canonical, the internal links, the sitemap and any redirects all point to it.

Why it matters for AI visibility

Google states that a page must be "indexed and eligible to be shown in Google Search with a snippet" to appear as a supporting link in AI Overviews or AI Mode, and that there are no additional technical requirements (Google's AI features documentation, last updated December 2025). Canonicalization decides which duplicate gets indexed. So a canonical conflict can decide which URL is even eligible to be shown as a source.

The index gate is real in practice. In OtterlyAI's 2026 experiment, Google's AI surfaces almost stopped citing two sites once Google deindexed them, while Microsoft Copilot kept citing them because Bing still held them. An engine cannot cite a URL its index does not hold.

What this does not mean: fixing a canonical does not make a page more likely to be cited. It removes a block. Whether the indexed URL then appears in an answer depends on retrieval and the model, which you do not control.

For engines outside Google, as of October 2026, ChatGPT, Perplexity and Claude do not document how they treat canonical tags. Treat the canonical as a Google and Bing index control, and check crawler access separately. Canonical problems are visible in a page crawl. Whether Google actually indexed your chosen URL is not, and needs Search Console.

Common misconceptions

"A canonical tag forces the engine to use my URL"

It does not. It is one signal among several. If the rest of your site points elsewhere, the engine can override it.

"A canonical tag is the same as noindex"

No. noindex asks an engine to keep a page out of the index entirely. A canonical asks it to fold duplicates into one indexed URL. Mixing the two on related pages sends contradictory instructions.

"Canonicals tell AI which page is my official source"

Only indirectly. A canonical shapes which URL a search index keeps. It does not tell a language model which facts to trust. The question of whether you can tell AI engines which page is your official source has a longer answer.

"Canonicals boost rankings or citations"

No evidence supports that. They decide which duplicate counts, nothing more.

Frequently asked questions

Should every page have a self-referencing canonical tag?

It is common practice and does no harm. A self-referencing canonical states that this URL is the preferred version, which helps when parameters or tracking codes create duplicate addresses. What matters more is that the declared URL resolves with a 200 status, is indexable, and matches the URL your internal links and sitemap use.

What is the difference between a canonical tag and a 301 redirect?

A 301 redirect sends users and crawlers to another URL, so the old URL stops serving content. A canonical tag leaves both URLs live and asks engines to index only one. Use a redirect when the duplicate should not exist. Use a canonical when both versions need to stay reachable, such as filtered or tracked URLs.

Can a wrong canonical keep my page out of Google AI Overviews?

It can keep that specific URL out. Google requires a page to be indexed and snippet-eligible to appear as a supporting link. If Google indexes a different URL as canonical, that other URL becomes the eligible one. If the canonical points to a broken or blocked page, the content may not be eligible at all.

Does ChatGPT read canonical tags?

OpenAI does not document how its crawlers treat canonical tags, as of October 2026. What OpenAI does document is that OAI-SearchBot controls eligibility for ChatGPT search, and that your host or CDN must allow its traffic. Check those access gates first, and keep canonicals consistent for the search indexes that do document them.

Can I use a canonical tag across different domains?

Yes, a cross-domain canonical is valid. It is often used when the same article is syndicated to a partner site and you want the original indexed. The same rule applies: it is a signal, so the receiving engine can still choose the other copy if other evidence points there.

Sources

This page is AI-assisted. It was written from a verified evidence base and checked against our editorial policy. Engine behaviour changes often; facts are as of October 21, 2026.