The homepage redirects by browser language and offers no link to choose a version

In the reportLocale auto-redirect blocks the no-language crawler

locale_auto_redirect_without_choice · Issue

Short answer

This appears when your homepage redirects a request to a language version chosen from its Accept-Language header, and the homepage served without that header has no ordinary link to that version. Google's crawler sends no Accept-Language header, so it may only ever see one version. Add plain links to every language version on the homepage.

Severity: High.

Why it matters

Google says that for pages which change by the visitor's language or country, it might not crawl or index all the content for different locales, because its crawler sends requests without an Accept-Language header. It recommends separate addresses for each locale, annotated with hreflang.

What to do

  1. Compare the recorded profile IDs, request headers and final addresses with the homepage’s recorded ordinary links to see which language destination lacked a choice link.
  2. Put ordinary links to every language version on the homepage, for example a language switcher made of a elements with href addresses.
  3. Consider suggesting a language with a banner in place of an automatic redirect, so that every version stays reachable at its own address.

What we read

We select at most two normalized language codes from the admitted homepage's hreflang declarations, using both its complete HTML and Link-header channels. The browser control follows the audit's smartphone or desktop collection profile. We request the homepage twice without Accept-Language and twice per selected code with that header added, without cookies on any request or redirect.

We report a stable HTTP redirect to a different address on the same site in both language requests when both complete control pages contain no ordinary link to that destination. The control must return the same complete, unchallenged content twice.

Example & evidence

Illustrative example — one audit, not a live result.

The homepage of bakery.example declares hreflang for de and fr. An audit requests it three ways, twice each:

GET https://bakery.example/                              200 (links: /recipes/, /about/)
GET https://bakery.example/   Accept-Language: de        302 -> /de/   then 200
GET https://bakery.example/   Accept-Language: fr        302 -> /fr/   then 200

The page served without a language header links to neither /de/ nor /fr/, so the site is reported. A possible correction: the homepage links to both versions; the redirects may stay, and nothing is reported.

Confirm & check again

Check for yourself: request your homepage with a command-line client, with and without an Accept-Language header, and check whether the page you get without the header links to each language version.

Check again: Full audit — Run a new audit. A complete clean recheck closes the same site finding only after every language that previously forced a destination is tested again. A language that disappears from the homepage's declarations remains unconfirmed; it does not prove a fix.

Limits & sources

Only the registered HTTPS homepage and at most two language tags selected from its own complete hreflang declarations are tested. Only language or language-region tags are supported; script tags and x-default are excluded. Missing declarations select no fallback language and leave this check unconfirmed.

Requests share one audit location. IP or country variation, JavaScript redirects and meta refresh are not tested. A chain longer than four redirects, an incomplete or unstable response, an audit-permission refusal or an exhausted request budget remains unconfirmed.

Hreflang annotations do not count as ordinary choice links. A redirect shared by requests with and without a language header does not fire this rule. Content that changes at the same address remains unconfirmed. These emulated headers do not identify a real crawler operator.

The destination test treats www and non-www names as the same site and does not compare schemes or ports. Every redirected request still requires HTTPS, public-address checks and the audit robots policy for its own origin. A refused HTTP target or a different site remains unconfirmed.

Confidence: High because the redirect responses and the links of the homepage are recorded as received, each request made twice, within the limit of two languages and one location.

Sources

  • Google, How Google crawls locale-adaptive pages — says Google might not crawl, index or rank all content of locale-adaptive pages, because the default addresses of its crawler appear to be based in the USA and it sends requests without an Accept-Language header, and recommends separate locale URL configurations annotated with hreflang. Checked 14 October 2026.

Editorial review: 3 November 2026