Free tool
Canonical checker
Most pages are reachable at several addresses: http and https, www and bare domain, with and without trailing slash or tracking parameters. Machines treat each as a separate document until you tell them otherwise. Enter a URL and this checker reports the canonical tag, whether it points at the page itself, and how your http to https and www redirects actually behave.
Why one address per page matters to answer engines
Every mention, link and crawl signal accrues to a URL. When the same content lives at four URLs, those signals split four ways, and the retrieval systems behind AI answers see four mediocre candidates where one strong page should stand. The canonical tag is your consolidation instruction: it names the single address that should collect the credit and appear in indexes, citations and previews.
A wrong canonical is quieter and worse than a missing one. Point a page at some other URL by template accident and you have formally asked machines to credit different content for your work. Nothing visibly breaks, traffic drifts, and the page you meant to promote fades from the sources engines quote. That silence is why this check belongs in any audit, even when everything looks fine in a browser.
How to read your results
A self-referencing canonical is the pass state: the page names its own preferred URL, exactly matching protocol, host and path. A canonical pointing elsewhere earns a warn, which is correct behavior on true duplicates such as parameter variants, and a defect anywhere else. A missing canonical leaves consolidation to the machine's guess, which usually works and occasionally picks the parameter-riddled variant to show the world.
The redirect checks tell the rest of the story. http should 301 to https in one hop, and your www choice should be enforced the same way, because each extra hop and each 302 where a 301 belongs weakens the consolidation message. When redirects and canonicals disagree, redirect to www while canonicalizing to bare domain for instance, machines receive two contradictory instructions and resolve the conflict without you.
What to fix first
Pick one host and one protocol, then enforce them with single-hop 301s at the edge. Next, make every indexable page emit a self-referencing canonical from the template so the rule survives future pages. Then hunt the contradictions this report surfaces, canonicals disagreeing with redirects, or pages canonicalizing to URLs that themselves redirect. Once your identity is consolidated, the interesting question becomes whether engines cite that one strong URL, and Reachroller tracks exactly that from $29 per month with a free 3-day trial.
Frequently asked questions
My canonical points to a different page. Is that always wrong?+
It is correct when the page is a true duplicate: a UTM-tagged variant, a print version, a session URL. It is a defect when a unique page canonicalizes elsewhere, because you are asking machines to ignore the page entirely. Check whether the target contains the same content; if it does not, the tag is misfiring and should self-reference instead.
Both http and https versions of my site load. How bad is that?+
Bad enough to fix this week. Two live protocols mean two crawlable copies of everything, split signals, and browsers or engines occasionally surfacing the insecure one. The fix is a single-hop 301 from every http URL to its https twin at the CDN or server level, which takes one rule and immediately halves your duplicate surface.
Should I use a canonical tag or a redirect?+
A redirect when users should never see the alternate URL, which covers protocol, host and retired paths. A canonical when the alternate must stay loadable, such as filtered or tagged variants of a page. The redirect is the stronger signal since machines have to obey it, so reserve canonicals for cases where both URLs genuinely need to serve content.
Do AI engines respect canonicals the way Google does?+
The assistants that answer live questions retrieve documents through search-grade indexing pipelines, and canonical consolidation happens in that pipeline before any answer is written. Gemini grounds on Google's index directly, and ChatGPT search relies on comparable infrastructure. You are maintaining one identity per page for the whole retrieval stack, and every engine downstream inherits it.
More free tools and reading
Tools find the gaps. Tracking closes them.
Reachroller asks the questions your buyers ask and shows which brands ChatGPT names. Three days free, no card.
Check my brand free