slugme.ioclean-url-slug-generator

URL Slugs for Arabic, French & Spanish: A Practical Guide

Accented and non-Latin text breaks most slug tools. Learn how to create clean, readable URL slugs for Arabic, French, and Spanish content the right way.

By 4 min read
URL Slugs for Arabic, French & Spanish: A Practical Guide

Most slug generators were built with English in mind. Feed them an accented French phrase or an Arabic headline and they often strip the meaning right out of your URL — turning "Diseño Web" into /dise-o-web or an Arabic title into an empty string.

If you publish in more than one language, you need slug rules that respect each script. Here's how to handle the three most common cases cleanly.

The core problem: naive cleaning destroys meaning

A typical slug function does something like "remove everything that isn't a–z or 0–9." That's fine for plain English. But applied to other languages it deletes the very characters that carry the meaning:

Diseño Web   → dise-o-web      (the ñ was deleted)
Café Crème   → caf-cr-me       (accents gone, words broken)
نصائح السيو  → (empty)          (the whole script removed)

The fix is to apply language-aware rules instead of one blunt filter.

French: convert accents, don't delete them

French uses accents heavily — é, è, ê, à, ç, ï, and more. The correct approach is transliteration: convert each accented letter to its base Latin form before cleaning.

é, è, ê, ë → e
à, â       → a
ç          → c
î, ï       → i
ô          → o
û, ù       → u

So the slug stays readable and keyword-relevant:

L'optimisation du référencement → /l-optimisation-referencement
Café & Crème                    → /cafe-creme

Technically this is done by normalizing the text (Unicode NFD) and removing the combining diacritical marks — then lowercasing and hyphenating as usual.

Spanish: handle ñ, accents, and tildes

Spanish has the same accent situation as French, plus the distinctive ñ. Convert each to its base letter:

á, é, í, ó, ú → a, e, i, o, u
ñ             → n
ü             → u

Examples:

Diseño Web en España → /diseno-web-espana
Guía de Programación → /guia-programacion

Note that ñ → n is the standard, readable convention for URLs even though they're different letters in Spanish — it keeps the slug ASCII-safe and predictable.

Arabic: preserve the script and strip the marks

Arabic is the trickiest case, for two reasons: it's written right-to-left (RTL), and it uses optional diacritics (Harakat) plus an elongation character (Tatweel).

The wrong move is to transliterate Arabic into Latin letters — that usually produces unreadable results. The right move is to keep the Arabic characters and only remove what doesn't belong:

  1. Strip Harakat (the vowel marks: َ ُ ِ ّ ْ …).
  2. Remove the Tatweel elongation character (ـ).
  3. Drop punctuation and symbols, but keep the Arabic letters intact.
  4. Replace spaces with hyphens.

The result is a clean, fully Arabic slug:

أفضل 10 نصائح في السيو → /أفضل-10-نصائح-السيو

Modern browsers and search engines fully support non-Latin (IRI) URLs, so an Arabic slug is valid, indexable, and far more meaningful to Arabic readers than a transliterated guess.

German: umlauts need two letters

German deserves its own note because the "just strip the accent" approach produces wrong words. The established convention is to expand umlauts:

ä → ae    ö → oe    ü → ue    ß → ss
Straße in München → /strasse-muenchen

Dropping the dots instead (ü → u) can turn one word into another — zählen (to count) becomes zahlen (to pay). See the German slug generator for the full rule set.

Cyrillic, Greek and letters NFD can't fix

The common Unicode trick — normalize to NFD and delete combining marks — works for é and ñ, but silently deletes letters that are not built from a base letter plus an accent: ß, ø, ł, æ, œ, đ, þ. A naive slugify turns "Łódź" into odz and "Ørsted" into rsted. The same function returns an empty string for Cyrillic or Greek titles.

A robust slug engine needs an explicit transliteration table on top of NFD:

Łódź Guide       → /lodz-guide
Æsir Ørsted      → /aesir-orsted
Привет мир       → /privet-mir
Καλημέρα κόσμε   → /kalimera-kosme

Don't forget stop words per language

Stop words differ by language. Removing English stop words from a French slug does nothing useful. Each language needs its own list:

  • French: le, la, les, un, une, des, et, de, du…
  • Spanish: el, la, los, las, un, una, y, de, del…
  • Arabic: في, من, على, إلى, و, ب, ل, مع…

A language-aware generator applies the right list automatically.

A quick comparison

Input Language Clean slug
Le guide de l'optimisation French /guide-l-optimisation
Straße in München German /strasse-muenchen
Diseño web en España Spanish /diseno-web-espana
أفضل 10 نصائح في السيو Arabic /أفضل-10-نصائح-السيو

Get it right automatically

Doing all of this by hand — per language, per script — is error-prone. Slugme.io ships with dedicated engines for English, French, Spanish, German and Arabic: it transliterates accents for Latin scripts (with German-style umlauts in the German engine), converts Cyrillic and Greek to Latin letters, preserves Arabic while stripping Harakat and Tatweel, flips the interface to RTL when needed, and applies the correct stop-word list for each language.

Pick your language, paste your title, and copy a clean slug — no broken characters, no empty strings. For the fundamentals, see what a URL slug is and the best-practices checklist.

Frequently asked questions

Does Google index URLs with Arabic or Chinese characters?

Yes. Google crawls and indexes Unicode URLs and displays them in readable form in search results. Its URL guidelines recommend using words in your audience's language.

Why does my Arabic URL look like %D8%A3%D9%81… when I paste it?

URLs are transmitted as ASCII, so non-ASCII characters are percent-encoded. Browsers decode them for display in the address bar, but some apps paste the encoded form. The link still works either way.

Should Russian URLs use Cyrillic or Latin letters?

Both work. Native Cyrillic matches how people search and reads naturally; transliterated Latin (privet-mir) is shorter when encoded and survives every system. Pick one convention and apply it site-wide. Slugme.io's English engine transliterates Cyrillic automatically.

Keep reading