Auto Language Redirect

Element Code: HR-024

What this check is flagging

This fires when your site auto-redirects visitors to a specific language or country version based on signals like their IP geolocation or their browser's Accept-Language header, instead of letting them land on the URL they requested. A French visitor asks for example.com/ and gets bounced to example.com/fr/. A German IP hitting example.com/en/ gets forced to example.com/de/. The redirect is usually a 302, sometimes a 301, and often it runs on every single URL, not just the homepage.

It feels helpful. The intent is "show people their language automatically." The execution is what causes the damage, because a crawler is not a person and does not carry the signals your redirect logic expects.

Why this wrecks international SEO

Googlebot crawls predominantly from US-based IP addresses and sends an Accept-Language header that is typically empty or set to en. Google has stated its crawler does not vary its Accept-Language header the way a browser does. So when your redirect logic keys off IP or language, Googlebot gets treated like a generic US English visitor and is funneled to one version. Every request for a French, German, or Japanese URL gets 302'd back to the English one before Google can render it.

The result is brutal and quiet: your non-default language versions barely get crawled, so they barely get indexed, so they never rank in their target markets. You built ten language versions and Google effectively sees one. I have watched sites with beautiful localized content pull zero organic traffic in-country because an auto-redirect strangled the crawl at the door.

There is a redirect-loop risk too. If the English version also has logic that tries to redirect based on a stale cookie or a misread header, you can create a chain or loop that returns a crawl error outright. And for users, forced redirects are genuinely annoying: a bilingual person who wants the English page cannot stay on it, and a traveler on foreign wifi gets a language they do not read.

Redirect versus the correct pattern

The correct pattern: serve requested URL, signal with hreflang Googlebot or user requests /fr/ 200 OK, French page no redirect hreflang links all versions to each other every version indexed

How to detect it

• Curl with different headers: the fastest check. Run curl -sI -H "Accept-Language: fr-FR" https://example.com/en/page and look at the status line. A 301 or 302 to a different language URL means an auto-redirect. Repeat with de-DE, ja-JP, and no header at all.
• Screaming Frog: crawl and check the Response Codes tab for unexpected 3xx on language URLs. You can also set a custom Accept-Language header in the crawl config and compare two crawls: one as en, one as another locale. Different destinations reveal the redirect.
• Search Console URL Inspection: inspect a non-default language URL. If Google reports it as a redirect or shows the wrong language in the rendered view, the auto-redirect is catching Googlebot.
• Server logs: a pile of 302s on localized URLs, especially clustered by the crawler's user agent, is a clear signal.
• A VPN spot check: hit the same URL from different countries and watch whether the URL in the address bar changes on its own.

How to fix it, step by step

1. Turn off the automatic redirect for crawlers, and ideally for everyone. The cleanest fix is to stop force-redirecting entirely. Serve whatever language version the URL asks for with a 200.
2. Pick a sensible default for the root. If someone lands on the bare domain with no language in the path, serve a default (often English or a country-selector page). Do not redirect deep localized URLs.
3. Offer a language switcher, not a forced jump. You can still detect a visitor's likely language and suggest it with a dismissible banner or a prominent switcher. Suggest, do not hijack. Google's own guidance is to let users select and to avoid automatic redirection based on perceived language.
4. Implement clean hreflang. Every language version references itself and every alternate with <link rel="alternate" hreflang="fr" href="...">, plus an x-default. This is how you tell Google which version serves which audience without moving anyone.
5. Make hreflang bidirectional and canonical-consistent. If A points to B, B must point back to A, and each page's canonical should be self-referential, not pointing at the default language.
6. Recrawl and re-inspect. Confirm localized URLs now return 200 for a crawler and that hreflang validates with no missing return tags.

Redirect method versus signal, compared

Do this, not that

What good looks like

Request /fr/, /de/, or /ja/ with any Accept-Language header and you get a 200 and the matching content, never a redirect to English. The bare domain resolves to a sensible default or a chooser. hreflang tags form a complete, reciprocal set with an x-default, and Search Console shows each language version indexed under its own URL in its own market. Users see a polite suggestion to switch languages, and they can ignore it and stay put.

FAQ