Can WordPress slugs use accents and non-English characters?

Yes. WordPress accepts any script in a slug. Accented Latin letters are converted to their plain forms, so café becomes cafe, while Cyrillic, Greek, Arabic, Chinese and other scripts are kept and stored percent-encoded. Google handles these URLs and recommends words in your audience's language.

What core does, script by script

Two steps run when a title becomes a slug. First, accented and special Latin characters are replaced with ASCII equivalents. Second, whatever is still outside ASCII is percent-encoded as UTF-8.

TitleStored slugWhat the browser shows
Crème brûléecreme-bruleecreme-brulee
Straße und Größestrasse-und-groesse on German-language sites, strasse-und-grosse otherwisethe same
Лучший кофе%d0%bb%d1%83%d1%87%d1%88%d0%b8%d0%b9-%d0%ba%d0%be%d1%84%d0%b5лучший-кофе
最高のコーヒー%e6%9c%80%e9%ab%98...最高のコーヒー

The German row shows that the conversion depends on the site language: core has locale-specific rules for German, Danish, Catalan, Serbian and Bosnian, so ö becomes oe on a German site and o elsewhere.

Percent-encoding is only the storage and transport form. Browsers display the decoded text in the address bar, and Google shows readable characters in results. The ugly version appears when the link is pasted somewhere that does not decode it: some chat apps, plain-text email, spreadsheets, log files.

What Google says about non-ASCII URLs

Google's URL structure guidance recommends using words in your audience's language in URLs, and gives non-English examples. It asks that non-ASCII characters be UTF-8 encoded, which is what WordPress does. So native-script slugs are fully supported; there is no need to romanise for Google's sake.

What a native-script slug gives you is the same mild benefit any descriptive slug gives: readers see words they understand in the result and in the address bar. It is not a strong ranking signal in any language.

Native script or transliteration?

Both are valid. Choose per site, not per post, so addresses stay consistent.

Native script suits sites written for one language community that reads URLs in the browser, where recognisable words in the address build trust.

Transliteration or English slugs suit sites whose links travel through email, messaging, print or tools that mangle encoded addresses, and teams who type URLs by hand.

Two practical points weigh toward shorter or transliterated slugs:

  • Length. The slug column holds 200 characters, counted after encoding. A Cyrillic letter uses six of them and a Chinese character nine, so around 33 Cyrillic letters or 22 Chinese characters fit, and longer titles are cut. See how long a slug can be.
  • Sharing. A 30-letter Cyrillic slug becomes 180 characters of %d0%... when an app fails to decode it, which looks suspicious and often wraps or breaks.

WordPress does not transliterate non-Latin scripts by itself. If you want luchshiy-kofe you type it in the slug field, or use a plugin that converts on save. Whichever you choose, write a short slug by hand; do not let a whole headline become the address.

Things to check on a multilingual site

  • Encoding consistency. Internal links should use the same form of each address. WordPress generates encoded links itself; hand-typed links in native script work too, since browsers encode them, but keep to one habit.
  • Case of the hex digits. Core stores encoded slugs with lowercase hex (%d0), and its redirect logic treats %D0 and %d0 as equal, so either form reaches the page.
  • Sitemap entries. URLs in an XML sitemap must be encoded and entity-escaped. WordPress and Hydrogen SEO output them that way; you can verify a file with the XML sitemap validator.
  • Language signals. A native-script URL does not tell search engines which language or region a page targets. That is the job of the content, the lang attribute and, for translated pages, hreflang.
  • Existing URLs. Do not convert a live site from one style to the other for tidiness. Every address would change and need a redirect.

Common questions

Will Google show my Cyrillic or Arabic URL correctly?

Yes. Google decodes UTF-8 URLs and displays the readable characters in results, as browsers do in the address bar.

Why does my slug turn into %d0%b1 when I paste it?

That is the percent-encoded form in which the address is stored and sent. It is the same URL; the app you pasted into simply did not decode it for display.

Does WordPress remove accents from slugs automatically?

For Latin letters, yes: é, ñ, ü and similar become plain letters when the slug is saved. Characters from other scripts are kept.

Can I mix Latin and non-Latin words in one slug?

Yes, for example a brand name followed by native-script words. Only the non-ASCII part is encoded.