Aizii Learn · AI Search

[← Back to Learn](https://aizii.ai/learn)

# Hreflang and international stores in AI search

Hreflang is a label that tells a search engine which language and region each version of a page is for, so the engine can point each shopper to the version that fits them. A store that sells in the United States, the United Kingdom, and Canada, in English and French, has several versions of the same product page: one price in dollars, one price in pounds, one page in French, and one page in English. Without hreflang, a search engine may show a shopper the wrong version: the French page to an English shopper, the pounds price to a dollar shopper, or the United States page to a shopper in Canada. An AI assistant that reads the wrong version will cite the wrong price, the wrong currency, and the wrong shipping terms. This guide covers how to label the versions, how to link them, and how to keep the facts straight per market. It does not promise that a label gets a page picked, and it does not cover translating the store. Hreflang connects pages that exist. It cannot create a translation that does not exist.

This is a developer-led job with one marketer check: the developer owns the tags, the sitemap entries, and the address structure, and the marketer owns the decision about which markets and languages the store serves. Agree the market list first, then add the tags, before you check anything.

## Where it goes wrong

International stores fail in ways that look fine to a shopper in the home market and break for everyone else.

-   Each language version has no separate address. The same address shows a different language based on a cookie or a browser setting, and the crawler, which usually comes from one country, sees only one version. Google recommends using different addresses for each language version rather than switching the language with cookies or browser settings.
-   The versions do not point back at each other. The English page links to the French page, and the French page does not link back. Google says that if two pages do not both point to each other, the tags will be ignored.
-   A version lists only some of the others. The English page lists the French page and forgets the Canadian page, or lists itself and forgets the fallback. Google says each language version must list itself as well as all other language versions.
-   The code is wrong. A page for British shoppers is tagged with a reserved code instead of the right one, or a region code stands alone with no language code. Google says the language must use an ISO 639-1 code and the region must use an ISO 3166-1 Alpha-2 code, and that reserved codes have no effect, so a page for British users should use en-GB.
-   The addresses are relative. A hreflang tag points to a path with no domain, or to an address with no transport method. Google says alternate addresses must be fully qualified, including the transport method, so a full https address, not a bare path.
-   The price and the currency do not match the market. The United Kingdom page shows a price in dollars, or the Canadian page shows the United States shipping terms. A model that reads the page will cite the price it sees, and the shopper will land on a page that charges them in the wrong currency.
-   The canonical tag crosses languages. The French page carries a canonical tag that points to the English page, telling the engine the English page is the representative version. Each version's canonical tag should point to a page in the same language, never across to another language version.

## The steps

**Step 1. List the markets and the languages before you tag anything.** The marketer lists every market the store serves and every language each market gets, and checks which pages exist in each one. A store that sells in the United States, the United Kingdom, and Canada, in English and French, needs a page list per market, not one global list. If a page exists in English and not in French, do not tag a French version for it. Hreflang can only connect pages that exist, and a tag that points to a page that does not exist will be ignored.

**Step 2. Give each version its own address.** The developer makes sure each language and region version of a page has its own address, with the language or region in the address where the platform supports it. Do not serve two languages from one address that switches on a cookie or a browser setting. Google recommends different addresses for each language version, and warns that a crawler that usually comes from one country may not find all the variations on a single address. Keep the address structure the same across versions, so the English, French, and fallback pages differ only in the language part of the address.

**Step 3. Tag every version on every version.** The developer adds a set of hreflang link tags to the head of each version, one tag per version including the version itself, and repeats the same full set on every version in the set. Each tag carries the language code and the region code where one applies, such as en-US, en-GB, and fr-CA, and the full https address of that version. Add an x-default tag that points to the fallback page for shoppers whose language matches no version. Check the codes against the vendor docs: the language uses an ISO 639-1 code, the region uses an ISO 3166-1 Alpha-2 code, and a region code cannot stand alone.

**Step 4. Keep the facts straight per market.** The marketer checks the price, the currency, the shipping terms, and the stock status on each version against that market's store data. The United Kingdom page shows the price in pounds and the shipping terms for the United Kingdom. The Canadian French page shows the price in Canadian dollars and the shipping terms for Canada. The product structured data on each version should carry the same price and currency the page shows. A model that reads the French page should cite the French price, not the price from another version.

**Step 5. Point each canonical tag at its own language.** The developer checks the canonical tag on every version and makes sure it points to a page in the same language, usually to the version itself. Do not point the French page at the English page, and do not point the Canadian page at the United States page. Hreflang says the versions are the same content in different languages. The canonical tag says which address represents one version. The two signals must agree, or the engine will treat the versions as duplicates instead of as a set.

**Step 6. Translate the main content, not only the template.** The marketer checks that each translated version carries the main content in the target language: the product description, the key facts, and the shipping and returns terms. Translating only the navigation and the footer is not a real version of the page. If a page cannot be fully translated, do not tag it as a translated version. Leave it untagged, or link to it from the language selector as a page that exists only in one language, and say so in words on the page.

**Step 7. Check the set after you publish it.** The developer checks the tags on a sample of pages in every version: each version lists itself and every other version, every address is fully qualified with https, every address loads, and every tag points back both ways. Check the Search Console international targeting reports where the platform shows them, and fix any version the reports flag. Add tags for a new page the same day it is published, and remove the tags of a removed version from every other version the same day.

## The machine-readable piece

This is a set of hreflang link tags for a hypothetical product page with four versions: English for the United States, English for the United Kingdom, French for Canada, and a fallback page for shoppers whose language matches no version. Replace the placeholders in square brackets with your own addresses. Each tag carries a language code and a region code where one applies, and the full https address of that version. Every version in the set carries this same set of tags, including a tag for itself. The x-default tag points to the fallback page.

```
<link rel="alternate" hreflang="en-US" href="https://[yourdomain.com]/products/ember-candle" />
<link rel="alternate" hreflang="en-GB" href="https://[yourdomain.com]/en-gb/products/ember-candle" />
<link rel="alternate" hreflang="fr-CA" href="https://[yourdomain.com]/fr-ca/products/ember-candle" />
<link rel="alternate" hreflang="x-default" href="https://[yourdomain.com]/products/ember-candle" />
```

Check the tags after you add them: open each version of the page, confirm the head carries the full set of tags, and confirm each address loads the right version in the right language and currency.

## What it looks like when it works

On the page, a shopper sees their language, their currency, and their shipping terms, with a language selector that lists the versions that exist. Under the surface, each version carries the full set of hreflang tags, each tag points to a full https address that loads, each version's canonical tag points to a page in the same language, and the sitemap agrees with the tags. The price and the currency on each version match that market's store data. A shopper in Canada who searches in French finds the French Canadian page, and a shopper in the United Kingdom who searches in English finds the United Kingdom page, each with the right price.

Say you run a hypothetical home fragrance brand selling candles and diffusers in the United States, the United Kingdom, and Canada, in English and French. The marketer lists the markets and checks that each product page exists in English for the United States, in English for the United Kingdom, and in French for Canada. The developer gives each version its own address, adds the full set of hreflang tags to every version including itself, and adds an x-default tag that points to the United States English page as the fallback. A shopper in Canada asks an assistant in French for a candle under a set price in Canadian dollars. The assistant can find the French Canadian page, check the price in Canadian dollars, and cite your store instead of a competitor's United States page with the wrong currency.

### The fundamentals

These mechanics are shared across every store and covered in full in their own guides:

-   •The llms.txt format and a complete sample, see our [guide to writing a good llms.txt file](https://aizii.ai/learn/ai-search/how-to-write-a-good-llms-txt-file)
-   •The markup AI systems read to verify facts, see our [product schema guide](https://aizii.ai/learn/ai-search/how-to-add-product-schema-markup-for-ai-search)
-   •The complete crawler rule reference, see our [robots.txt guide for AI search](https://aizii.ai/learn/ai-search/how-to-write-robots-txt-for-ai-search)

## What this guide does not cover

This guide covers hreflang and international stores in AI search: the market list, the address structure, the tags, the per-market facts, the canonical rule, and the check routine. It does not cover translating the store, choosing a market, setting prices per market, or the transaction layer that lets an agent complete a purchase. It does not cover the sitemap strategy for a large catalog in full, or the product feed build. Those are separate jobs, and this guide is the map for the international pages, not the whole trip.

## Check your own store now

Run the free scan. One click, no signup, and it shows whether AI crawlers can read your store today: aizii.ai/aeo-scan

[Run the free scan](https://aizii.ai/aeo-audit)

## Sources

-   Google Search Central, Localized versions of your pages (specific): [https://developers.google.com/search/docs/specialty/international/localized-versions?visit\_id=639021532213139460-1305425403&rd=2](https://developers.google.com/search/docs/specialty/international/localized-versions?visit_id=639021532213139460-1305425403&rd=2)
-   Google Search Central, Managing multi-regional and multilingual sites (specific): [https://developers.google.com/search/docs/specialty/international/managing-multi-regional-sites?hl=en](https://developers.google.com/search/docs/specialty/international/managing-multi-regional-sites?hl=en)
-   Google Search Central, Product structured data (specific): [https://developers.google.com/search/docs/appearance/structured-data/product](https://developers.google.com/search/docs/appearance/structured-data/product)
-   Google Search Central, Optimizing your website for generative AI features on Google Search (specific): [https://developers.google.com/search/docs/fundamentals/ai-optimization-guide](https://developers.google.com/search/docs/fundamentals/ai-optimization-guide)
-   Schema.org, Product type (specific): [https://schema.org/Product](https://schema.org/Product)

Note: links marked specific point to the page that supports the claim in this guide. Links marked general docs point to the vendor's general docs section for that topic, not to a page that proves every step. Check the vendor docs for the current path before relying on a menu name or click path in this guide.
