Hreflang checker and hreflang tester tools, what they validate and the en-GB mistake British sites make most

An hreflang checker reads one page, pulls out its alternate language and region declarations, and tells you three things: whether every code is a real one, whether each page it names points back at the page you checked, and whether a fallback has been declared for visitors the set does not cover. Those three are the whole job. Almost every hreflang fault found on a real site is one of them, and the commonest of the three on a British site is the return link, because somebody added a new market and only edited the new pages.

Google reports 140 searches a month in the United Kingdom alone for "hreflang checker" and 30 for "hreflang tester", harvested on 22 October 2026 against the United Kingdom target on its own.

What the tags are for, in Google's own words

Google's documentation on localized versions, read at developers.google.com on 22 October 2026, states the purpose plainly: "If you have multiple versions of a page for different languages or regions, tell Google about these different variations. Doing so will help Google Search point users to the most appropriate version of your page by language or region." The same page names the case this is for: "If your content has small regional variations with similar content, in a single language. For example, you might have English-language content targeted to the US, GB, and Ireland."

That is the British situation almost exactly. A company in Manchester selling the same thing to Britain, Ireland and the United States has three pages of nearly identical English, and the decision of which one a searcher in Dublin should land on is the decision these tags exist to inform. For the wider architecture these tags sit inside, see multilingual SEO architecture, hreflang and localisation that holds up.

The three checks every tester runs

• Are the codes valid. The language part comes first and the optional region part second. So en-GB is English as served to the United Kingdom, and en-IE is English as served to Ireland. A region on its own is not a valid value, which is why uk as a standalone code does nothing. Worth noting while you are in there: the country element of the United Kingdom's two letter code is GB, not UK, and a tester will mark gb as valid and uk as not.
• Do the links return. If the British page names the Irish page, the Irish page has to name the British page. Google's documentation is explicit that newly expanded language pages should be linked bidirectionally to the originating pages. A one way declaration is the fault a checker catches fastest and the one a human reviewer almost never spots.
• Is there a fallback. The documentation recommends "adding a fallback page for unmatched languages, especially on language/country selectors or auto-redirecting home pages" using the x-default value. Without it, a visitor in a market you did not list has no declared destination.

The three places the tags can live, and why that matters to the checker

The declarations can sit in the head element of the page, in the HTTP response headers, or in your sitemap. Google accepts all three. A checker that only reads the head element will report a correctly configured site as having no declarations at all if that site put them in the sitemap, and this is the single commonest false alarm in this category.

Where the tags live	When to use it	What a checker needs to do
The head element of each page	The default, and the right choice when you have no sitemap control or cannot set headers	Fetch the page and parse the head
HTTP response headers	Non HTML files, such as a PDF price list served in three markets	Read the response headers, which many browser tools do not show by default
The sitemap	Large sites, because one file holds the whole set and one edit updates every page in a group	Fetch and parse the sitemap, which single page checkers do not do

Pick one method per page group and stay with it. Declaring the same group in two places is how two sets drift apart, and the drifted one is usually the sitemap because nobody opens it.

How to check a British site in ten minutes

• List the page groups, not the pages. A group is one piece of content in all its market versions. Your pricing page in three markets is one group. Checking page by page is how you miss that the group is incomplete.
• Take the British page of one group and read its declarations. Note every address it names.
• Open each of those addresses and check it names the British page back. This is the return link test done by hand, and doing one group by hand teaches you what the automated report is telling you.
• Check the set includes the page itself. A group should declare every member including the one you are standing on. Self reference missing is a common output of templates that build the list from the other markets.
• Check x-default exists somewhere in the group. Usually on the global or selector page.
• Then run a crawler over the whole site and look only for groups where the count of declarations is not the same on every member. Unequal counts inside one group is the report that finds real faults.

Three things an hreflang checker will not tell you

A clean validation report means the markup is consistent. It does not mean the pages are right.

• Whether the pages genuinely differ. Three identical pages with different declarations give a searcher nothing to gain from landing on one rather than another. The regional variation has to be real: the currency, the delivery promise, the VAT treatment, the contact number, the spelling.
• Whether a redirect is undoing the work. A site that automatically sends every visitor to the version matching their location can override the whole point of the declarations. Offer the choice rather than forcing it.
• Whether the markets should be separate pages at all. Two pages for Britain and Ireland that say the same thing in the same language are two pages competing for one set of queries. One page with a clear note on Irish delivery and pricing is often the stronger answer. The wider decisions about which markets to treat separately are in international marketing explained, the four decisions and what to adapt per market.

Questions about hreflang checkers

Should a British site use en-GB or just en

Use en-GB when you have a separate page for Britain and at least one other English market, because the region part is what distinguishes them. Use plain en when you have one English page serving everybody, since claiming a region you have not built a page for adds nothing. The code for the United Kingdom is GB, and uk is not a valid region value in this markup.

Why does my hreflang checker say the tags are missing when I can see them

Three usual reasons. The tags are in your sitemap or your response headers and the tool only reads the head element. The tags are injected by JavaScript after load, so a raw fetch does not see them. Or the tool followed a redirect and validated a different address from the one you meant.

What does a return link error actually mean

Page A names page B as an alternate, and page B does not name page A back. Google's guidance is to link newly expanded language pages bidirectionally with the originating pages, so a one way declaration is an incomplete instruction. It almost always happens because a new market was added and only the new pages were edited.

Do I need x-default

Add it where you have a selector page or a home page that routes people, which is what the documentation recommends it for. It names the destination for a visitor whose language and region you have not listed, and on a site serving three English markets out of a dozen countries' worth of traffic, that is most visitors.

Can hreflang fix duplicate content between my British and American pages

It is the right signal for the situation rather than a repair for it. The declarations tell Google the pages are regional variants of one thing rather than copies, which is what you want recorded. Where the two pages genuinely say the same thing, the stronger move is one page with the regional detail inside it.

Which hreflang checker should I use

For one page, any validator that reads the head element, the response headers and the sitemap, and says which of the three it found. For a whole site, a crawler that groups the declarations and shows you where one member of a group disagrees with the rest. The grouped view is the one that finds faults.

The other market signal, and the one nobody validates

Language and region declarations tell a search engine which version of a page to show a person in Dublin or Leeds. There is a second surface now making the same decision with none of this markup. When somebody asks an assistant for a supplier in Britain, it assembles an answer and names companies, and what it draws on is whatever has been published about them rather than any tag you control. British businesses are routinely handed answers built from American sources, which is a specific and measurable problem rather than a general one. Knowing what the assistants currently return for your own market, with your name left out of the question, is the equivalent reading for that surface. AI Knows Us asks those buying questions across the main assistants and records who gets named and which sources the answer came from.