Multilingual SEO & GEO
Proxylang translates your pages on the server, so search engines and AI tools can index every language you offer. Then it measures the results and gives you a score.
Included from the Start plan. Set up with a single DNS record, no code changes.
Translated on the server
Every translated page is built on our servers before it is sent. The title, description, language tags and full text arrive already translated, so search engines can index each page right away.
<html lang="en">
<title>Waterproof Hiking Boots</title>
<meta name="description" content="Built for wet trails.">
<h1>Waterproof Hiking Boots</h1>Visitors can read the translation, but search engines only ever see the original English.
<html lang="ko">
<title>방수 등산화</title>
<meta name="description" content="젖은 등산로를 위해.">
<link rel="canonical" href="/ko-kr/boots">
<link rel="alternate" hreflang="ko-KR" href="/ko-kr/boots">
<link rel="alternate" hreflang="x-default" href="/boots">
<h1>방수 등산화</h1>A translated page with its own address, linked to every other language. Korean is shown here, but every language you add works the same way.
Measured
We load your translated pages the same way search engines do and grade the results. You get one clear number, the details behind it, and a weekly history so you can watch it improve.
Example figures for a fictional store.
Automatic
There are no settings to manage and no tags to keep up to date. If you add another language next year, all of this is set up for it automatically.
URLs
Korean lives at /ko-kr/pricing, Japanese at /ja-jp/pricing. Same domain, so every language inherits the authority you already have.
Each page links to every other language, itself, and an x-default — in the HTML and the HTTP header. Any hreflang your site already emits is stripped first, so the two can never disagree.
The Korean page is canonical to itself. This is the one detail that stops Google reading your languages as duplicates.
html lang, og:locale and content-language on every page. Google, Bing and Baidu each read a different one. You get all of them.
In the index
Each page is translated on the server before it is sent. Nothing needs to run in the browser for it to be indexed.
Titles, meta descriptions and social preview text are translated with the page, so your listing in search results appears in the right language.
Image descriptions are translated as well, which helps your images appear in image search in each language.
Product details, articles, FAQs and other structured data are translated, while prices, dates and links stay exactly as they were. Right-to-left languages display in the correct direction.
Discovery
One combined multilingual sitemap, plus a separate sitemap for each language. Large sites are split into a sitemap index automatically.
Your existing rules carry through untouched. Only the sitemap entries are replaced with the full multilingual set.
Bing, Yandex, Naver and others are notified as soon as new translations go live. Connect Google Search Console once and your sitemaps are submitted for you on a schedule.
Answer engines
The crawlers behind ChatGPT, Claude and Perplexity receive the same translated pages that Google does.
AI crawlers are welcomed in robots.txt, and a summary file at /llms.txt lists your languages and sitemaps for them.
Not just Google
Each search engine looks for different language signals. Google wants one tag, Bing another, and Naver mostly wants a Korean sitemap. Your pages include the full set, so no market gets left out.
| Engine | hreflang | Content-Language | og:locale | IndexNow | Sitemap | robots.txt |
|---|---|---|---|---|---|---|
Google hreflang is the signal that counts. Monitored through Search Console. | hreflang: Primary signal | Content-Language: Ignored | og:locale: Read | IndexNow: Ignored | Sitemap: Read | robots.txt: Read |
Bing Reads the Content-Language header far more than hreflang. Takes IndexNow. | hreflang: Weak | Content-Language: Primary signal | og:locale: Read | IndexNow: Read | Sitemap: Read | robots.txt: Read |
Yandex Reads every signal on this list. | hreflang: Read | Content-Language: Read | og:locale: Read | IndexNow: Read | Sitemap: Read | robots.txt: Read |
Yahoo Japan Powered by Google, so it behaves like Google. | hreflang: Read | Content-Language: Ignored | og:locale: Read | IndexNow: Ignored | Sitemap: Read | robots.txt: Read |
Naver Wants a Korean sitemap and a manual registration. We generate both and walk you through it. | hreflang: Ignored | Content-Language: Ignored | og:locale: Ignored | IndexNow: Read | Sitemap: Weak | robots.txt: Ignored |
Baidu Content-Language is the signal that counts. A .cn domain also needs an ICP licence. | hreflang: Ignored | Content-Language: Primary signal | og:locale: Ignored | IndexNow: Ignored | Sitemap: Weak | robots.txt: Ignored |
Applebot, DuckDuckBot, Daum, Kakao, Sogou, Seznam, Petal, Mojeek and Qwant are recognised as well. The major crawlers are verified as genuine before they are served, so fake bots posing as Google cannot use up your translation quota.
Search engines that serve one country get that country's version. Naver and Daum receive the Korean pages, Baidu and Sogou the Chinese ones. Google and Bing crawl every language, so they can index all of them.
Plans
Search engines can only index what the server sends them. That is the difference between the plans.
Script tag
Translation runs in the visitor’s browser. Great for readers, but search engines still see your original language.
Hybrid
Search engines and AI crawlers get translated pages from the server, while your visitors keep the instant widget. Titles, descriptions and headings are always translated, and body text is translated in order of importance.
Full proxy
Every request is served fully translated from the server. The page your reader sees is the page Google stores.