Multilingual SEO services

Run a Dutch keyword list through a translation engine and you get French words nobody types. In every language people search with their own synonyms, their own question forms and their own balance between informational and commercial searches. Sometimes a whole topic does not exist in one language and is the largest category in another.

So Traze researches demand afresh per language: search volumes, People Also Ask questions, the questions AI assistants receive in that language and the competitors currently ranking. The result is a cluster map per language that overlaps where that makes sense and diverges where the market diverges. See how the content strategy builds clusters per language.

A translated page keeps the structure and arguments of the source language, even when they do not convince in the target language. Transcreation starts from the question in the target language and writes an answer to it, with the facts, tone of voice and positioning from your brand profile. The result is a page that reads as naturally in French as your Dutch original. Read how AI content creation writes in your voice, in other languages too.

Traze writes in the target language and works with a brand profile per language, so forms of address, terminology and examples are right. Where you want it, every page passes a native reviewer before it enters the approval queue. Every page gets a citability score per language, so you can see whether it also stands a chance in AI answers. See how the approval flow and the native check work.

Hreflang is the mechanism that tells Google which language version belongs to which visitor. It is fragile: every language version must reference all the others and itself, language codes must be valid, x-default must point to a deliberate fallback and canonicals must never point from one language to another. One error in the template copies itself to every page. See how technical optimization finds hreflang and canonical errors.

Traze generates hreflang and canonicals from your site's language registration and checks them on every publish. On existing sites we start with a scan of the whole cluster; the Tech Scan shows you for free whether your current hreflang implementation holds up. Scan your site for hreflang and canonical errors for free with the Tech Scan.

Every CMS handles languages differently. WordPress works with WPML or Polylang, which store each translation as its own post and emit hreflang themselves. Storyblok translates at field level within one story and lets the frontend decide the language routes. Shopify uses Translate & Adapt with language variants per market. Traze publishes into the structure your CMS already uses, so no second translation system grows beside it. Read our approach to WordPress SEO with WPML or Polylang.

Per CMS we watch for the pitfalls: duplicate hreflang tags when plugin and theme both emit them, untranslated slugs, language variants without their own metadata and sitemaps missing one language. Those errors are small but structural, and harmful precisely because of that. See how Traze works with field-level translations in Storyblok.

Most multilingual sites start tidy and grow lopsided: new pages appear only in the main language, updates never reach the other languages and orphan translations pointing at pages that no longer exist creep in. Google sees a site that is half finished in French, and judges it accordingly.

Traze treats language variants as one set. A new page in the calendar automatically gets its variants in the agreed languages, a content update is carried through in every language and the dashboard shows per language which pages are missing or lagging. Coverage stays even while the site grows. Read how this works for online stores with large catalogues.

AI assistants answer questions in the language they are asked in, and draw their sources largely from that language too. A brand that is mentioned everywhere in Dutch can be invisible in French, simply because the French pages do not answer the question directly or do not exist. See the example case of a fashion boutique in four countries.

We measure per language and per engine whether you are mentioned, write language versions that answer directly with clear entities and structure, and track which pages in which language earn citations. For brands in several countries we combine that with the country structure from international SEO. Read more about international SEO for brands in several countries.