Cross-lingual search
Type “sneakers”. Find “tênis”. And “zapatillas”. And “baskets”.
One shopper. One catalogue. Any language. AI Search matches meaning across languages, so a query in one language finds products described in another — no translation layer, no synonym list to maintain.
Install free from WordPress.org →10,000 embeddings/month free. No credit card. No OpenAI account.
Same query. Three catalogues. One search that gets it.
Below are real cross-lingual matches AI Search finds today. The shopper types in the language they think in; your catalogue keeps its own language. Both work.
🇵🇹 Portuguese catalogue
Tênis Nike Air Max
Sapatilhas Adidas Ultraboost
Ténis Puma Suede
🇬🇧 English catalogue
Ivory silk wedding guest dress
Blush midi wrap dress
Formal linen jumpsuit
🇮🇹 Italian catalogue
Cappotto in lana merino
Piumino lungo da donna
Giacca termica invernale
🇪🇸 Spanish catalogue
Set de velas aromáticas
Neceser cuero regalo mujer
Pulsera plata día de la madre
🇩🇪 German catalogue
Sportkopfhörer bluetooth
In-Ear Ohrhörer wasserdicht
Kabellose Fitness-Kopfhörer
🇬🇧 English catalogue
Leather wallet with initials
Whiskey tasting gift set
Personalised barbecue tool set
How this actually works
AI Search converts every product into a vector that captures its meaning, not its words. The query gets the same treatment. Because the model was trained on 100+ languages, the vector for “sneakers” (English) sits right next to the vector for “tênis” (Portuguese) — they mean the same thing, so they match. Your catalogue keeps its language. Your shoppers keep theirs. Both work.
This is not translation
Most “multilingual search” plugins do one of two things: they force one search per language (English query only finds English content, Portuguese query only finds Portuguese) or they run automatic translation on the query and hope for the best. Both approaches miss half the value.
AI Search does neither. The embedding model works at the level of meaning, not the level of words. There is no translation layer between “prezent dla taty” and “leather wallet”. They just share a mathematical neighbourhood in vector space. No synonyms to maintain, no per-language index to build.
Where cross-lingual matters
🌍
EU stores selling to expats
A store in Lisbon whose catalogue is Portuguese, but half the customers are British expats searching in English. AI Search bridges the gap without a second index.
🌐
WPML and Polylang sites
You already run separate copies of each product per language. AI Search creates one embedding per copy — so a query in any language finds the closest match in every language you carry.
🏖️
Tourist-heavy catalogues
Souvenir shops, restaurants, activities: your inventory is described in the local language but visitors search in theirs. Every missed match is a lost sale.
🌱
B2B wholesale, one catalogue
A supplier with a single English catalogue serving buyers across LATAM, DACH, and Southern Europe. Each buyer searches in their own language and lands on the same SKUs.
Compatible with the plugins you already use
AI Search does not replace WPML or Polylang. It runs on top: your existing per-language content stays exactly as it is, and AI Search adds semantic matching across languages for the queries WordPress default search can never reach.
Add cross-lingual semantic search to your store.
Free tier: 10,000 embeddings/month, unlimited free searches. No credit card, no OpenAI account, no synonym lists to maintain.
Install free from WordPress.org →