Canonical Tag and Noindex Together

Can you use canonical and noindex together?

Pick one: do not put a noindex rule and a rel=canonical that points to a different URL on the same page. The two signals ask for opposite outcomes, so the result depends on which one Google decides governs, and that is not something a site owner can control. If the goal is to consolidate a duplicate into a preferred URL, use the canonical alone. If the goal is to keep the page out of search results, use noindex alone.

The one pairing that is generally considered harmless is noindex with a self-referencing canonical, where the canonical names the URL it sits on. Nothing is being consolidated anywhere else, so there is nothing for noindex to contradict, and noindex decides whether it is indexed.

Why rel=canonical and noindex conflict

The conflict comes down to what each signal asks for. A cross-page canonical says: "this page is a duplicate; fold its signals into the target URL and index that target instead." A noindex rule says: "do not show this page in search results." The first treats the URL as a contributor to another URL's standing. The second takes it out of consideration. Google cannot honor both goals cleanly on one URL.

Here is the pattern in markup, a noindex robots meta tag beside a canonical that points elsewhere:

<meta name="robots" content="noindex">
<link rel="canonical" href="https://www.example.com/preferred-page/">

Read literally, the markup says "don't index me" and "my signals belong over there" at the same time. Google's documentation already rules out the idea of using noindex as a canonicalization tool within a site: "We don't recommend using noindex to prevent selection of a canonical page within a single site, because it will completely block the page from Search" (Google Search Central).

What Google does with the conflict, and why its statements vary

The two signals do not carry equal weight. Google follows noindex as a rule once it has crawled the page, while it calls rel=canonical "a hint, not a rule." Gary Illyes described the noindex side in 2020: "something with noindex will never reach the serving index, but we will have the fetched copy for things like link graph calculation." The common reading is that noindex prevails and the consolidation the canonical asked for may not happen; the fetched copy is a narrow technical detail, not a promise that the canonical will do its job.

Anyone researching this finds Google staff quoted on more than one side, and the quotes are real. In 2018 John Mueller called the pair "very contradictory pieces of information for us" and said "We'll generally pick the rel=canonical and use that over the noindex." In a 2021 office-hours session he said "You can also do both of them," adding "maybe some of the signals we can just forward along." In 2024 his advice on Reddit was "I'd just pick one," although one account of that exchange places the remark in a discussion of noindex versus followed links rather than canonicals.

The useful conclusion is not which quote is current. Three answers from the same person across six years describe an outcome nobody outside Google can predict, and nothing in Google's documentation recommends the combination.

A worked example: filtered category URLs

Faceted category templates are a common source of the pair. Suppose /shoes/?color=red and /shoes/?color=red&sort=price both load filtered versions of /shoes/, and the template writes this into every filtered URL:

<!-- on https://www.example.com/shoes/?color=red -->
<meta name="robots" content="noindex">
<link rel="canonical" href="https://www.example.com/shoes/">

Whoever built the template wanted two things: keep filtered URLs out of search, and credit /shoes/ with any links those URLs attract. The markup cannot promise the second. The fix is to decide which goal matters for these URLs:

• The filtered pages are near-duplicates of /shoes/. Remove the noindex and keep the canonical. If Google accepts the hint, the filtered URLs show up in the Page indexing report as "Alternate page with proper canonical tag," the expected state rather than an error.
• The filtered pages are not duplicates and should not appear in search at all. Keep the noindex and remove the cross-page canonical, or replace it with a self-referencing one.
• The filters generate an effectively endless set of URLs that eat crawl activity. A robots.txt disallow saves the crawling, but Google then reads neither the noindex nor the canonical on those URLs. Choose it knowingly, as a crawl control, not as a way to make either signal work.

When the pair hides in headers, robots.txt, or JavaScript

The conflict is not always two tags sitting next to each other in the <head>. Three variations make it harder to spot:

• The X-Robots-Tag header. The X-Robots-Tag HTTP response header carries the same rules as the robots meta tag, and a canonical can also travel as an HTTP Link header. A server or CDN can send noindex in a header while the template writes a cross-page canonical into the HTML. View Source shows only the canonical; the noindex is visible only in the response headers.
• robots.txt. A URL disallowed in robots.txt is never fetched, so Google reads neither its noindex nor its canonical. Google's guidance is "Don't use the robots.txt file for canonicalization purposes," and a blocked URL can still be indexed without its content when other pages link to it, which surfaces as "Indexed, though blocked by robots.txt."
• JavaScript. Google says that when it finds noindex it "may skip rendering and JavaScript execution" (JavaScript SEO basics). A canonical injected by script, or a script written to remove the noindex, may never run. The initial HTML is what decides.

Checking the headers takes one command. A response like the second line below, on a page whose HTML carries a canonical to another URL, is the same conflict in a different place:

curl -sI "https://www.example.com/shoes/?color=red" | grep -i '^x-robots-tag'
x-robots-tag: noindex

The mirror image: a canonical that points at a noindexed page

The same contradiction can span two pages. URL A carries a canonical to URL B, and B carries noindex. A asks Google to index B in its place; B asks not to be indexed at all. The canonical nominates a page that has taken itself out of the running, so Google has no clean target and may pick its own canonical or leave the content poorly represented. Screaming Frog reports this as a "Non-Indexable Canonical." The fix sits on the target: every canonical target should return a 200, be indexable, and not be blocked by robots.txt.

hreflang clusters raise the stakes. Gary Illyes said in 2022 that if Google treats the pages of a language cluster as duplicates, "one noindex can affect the whole cluster," so a noindex placed on one locale is not necessarily contained to that locale.

What to do instead

The fix follows from the two goals being incompatible:

• To consolidate a duplicate page into a preferred one: use a cross-page canonical alone, or a 301 redirect if the duplicate does not need to stay reachable. Do not add noindex to the duplicate; the canonical is the mechanism meant to handle the consolidation.
• To keep a page out of the index entirely: use noindex alone, keep the page crawlable so Google can read it, and drop the cross-page canonical. A self-referencing canonical can stay.
• To stop crawling of an endless URL space: use robots.txt, accepting that neither noindex nor a canonical on those URLs will be read.
• Do not use noindex to steer canonical selection. Adding noindex to a URL in the hope of nudging Google toward a different canonical is not a reliable technique; it removes the page from Search rather than influencing which URL represents it.

How to find pages with both a canonical and a noindex

The conflict usually comes from systems rather than a single decision: plugins or templates that emit both tags on paginated or faceted URLs, or a developer adding noindex to "fix" a duplicate whose canonical was already the intended fix.

To find affected pages, crawl the site and check each URL's HTML head and HTTP response headers for a robots meta noindex (or an X-Robots-Tag noindex) alongside a rel=canonical pointing to a different URL. Crawlers such as Screaming Frog and Sitebulb report meta robots and X-Robots-Tag separately and flag the combination. In Search Console, the Page indexing report's "URL marked 'noindex'" list can be compared against the crawl's list of pages that canonicalize elsewhere; any URL on both lists has the conflict. The URL Inspection tool confirms it one URL at a time, showing the noindex status next to the user-declared and Google-selected canonical.

Frequently asked questions