slugme.ioclean-url-slug-generator

Slugify Libraries Compared: JS, Python, PHP & Ruby

We ran 10 slugify libraries on the same tricky inputs (ß, Ł, Cyrillic, Arabic, apostrophes). Real outputs, the gotchas, and which library to pick.

By 6 min read
Slugify Libraries Compared: JS, Python, PHP & Ruby

Every web framework eventually needs to turn "The Ultimate Guide to Café Culture!" into ultimate-guide-cafe-culture. There are dozens of libraries for it, and their READMEs all promise "clean, URL-safe slugs". The interesting part is where they disagree — and you only find that out with awkward input.

So we tested them. Same nine inputs, ten libraries, four languages, real output copied from the terminal. No library authors were consulted; these are the defaults you get after npm install, pip install, composer require or gem install.

The test inputs

Each input targets a specific failure mode we have seen break real sites:

Input What it tests
Straße & Café German ß (doesn't decompose under Unicode NFD) and &
Hello World! The baseline — every library should pass
Привет мир Cyrillic script
Łódź Guide Polish Ł (another letter NFD can't fix)
أفضل نصائح السيو Arabic script
影師嗎 Chinese characters
Don't Stop Apostrophe inside a word
I ♥ Dogs Symbol with a meaning
Größe Über German umlauts (ö, ü)

JavaScript results

Tested: slugify (simov), @sindresorhus/slugify, transliteration and github-slugger.

Input slugify {lower, strict} @sindresorhus/slugify transliteration github-slugger
Straße & Café strasse-and-cafe strasse-and-cafe strasse-cafe straße--café
Привет мир privet-mir privet-mir privet-mir привет-мир
Łódź Guide lodz-guide lodz-guide lodz-guide łódź-guide
أفضل نصائح السيو afdhl-nsaeh-alsyw afdl-nsayh-alsyw fdl-nsy-h-lsyw أفضل-نصائح-السيو
影師嗎 (empty) (empty) ying-shi-ma 影師嗎
Don't Stop dont-stop dont-stop don-t-stop dont-stop
I ♥ Dogs i-love-dogs i-love-dogs i-dogs i--dogs
Größe Über grosse-uber groesse-ueber grosse-uber größe-über

What stands out:

  • slugify keeps the original case by default. Without options, slugify('Hello World!') returns Hello-World! — capital letters and the exclamation mark. Always pass { lower: true, strict: true }. With locale: 'de' it produces groesse-ueber and even turns & into und.
  • @sindresorhus/slugify has the best defaults of the group: lowercase, German-style umlauts, & → and, ♥ → love. It is ESM-only, and ships slugifyWithCounter(), which returns foo-bar, foo-bar-2, foo-bar-3 for repeated input.
  • github-slugger is not a URL slugger. It reproduces GitHub's heading anchors, so it keeps Unicode and can produce double hyphens. Use it for Markdown heading IDs, not page URLs.
  • Arabic transliteration is lossy everywhere. afdhl-nsaeh-alsyw is not something an Arabic reader would ever search. For Arabic content, a native-script slug is usually the better choice (more on that in URL slugs for non-Latin languages).

Python results

Tested: python-slugify 8.0.4 and Django's slugify (4.2), with and without allow_unicode.

Input python-slugify Django Django allow_unicode=True
Straße & Café strasse-cafe strae-cafe straße-café
Привет мир privet-mir (empty) привет-мир
Łódź Guide lodz-guide odz-guide łódź-guide
أفضل نصائح السيو fdl-nsyh-lsyw (empty) أفضل-نصائح-السيو
影師嗎 ying-shi-ma (empty) 影師嗎
Don't Stop don-t-stop dont-stop dont-stop
Größe Über grosse-uber groe-uber größe-über

What stands out:

  • Django's default silently deletes letters. slugify('Straße') is strae and slugify('Łódź') is odz. Any Cyrillic, Arabic or Chinese title produces an empty slug — which, if you use the slug as a URL or unique key, is a production bug waiting to happen. Django only promises ASCII by stripping; it does not transliterate.
  • python-slugify transliterates nearly everything, but splits don't into don-t. Fix that with replacements=[["'", ""]]. It also supports max_length with word_boundary=True, stopwords, and custom separator.
  • Neither does German umlauts by default (grosse-uber). In python-slugify, add replacements=[['ü', 'ue'], ['ö', 'oe'], ['ä', 'ae']].

PHP results

Tested: Laravel's Str::slug (illuminate/support 13), cocur/slugify 4.7 and Symfony's AsciiSlugger (symfony/string 8.1, with the intl extension).

Input Laravel Str::slug cocur/slugify Symfony AsciiSlugger
Straße & Café strasse-cafe strasse-cafe Strasse-Cafe
Привет мир privet-mir privet-mir Privet-mir
Łódź Guide lodz-guide lodz-guide Lodz-Guide
影師嗎 (empty) (empty) ying-shi-ma
Don't Stop dont-stop don-t-stop Don-t-Stop
Größe Über grosse-uber groesse-ueber Grosse-Uber

What stands out:

  • Symfony keeps case. Call ->lower() on the result: $slugger->slug($title)->lower(). Passing a locale (new AsciiSlugger('de')) gives groesse-ueber.
  • Laravel takes a language argument. Str::slug('Größe Über', '-', 'de') returns groesse-ueber. It also converts @ to at by default (hello@example.com → hello-at-examplecom).
  • cocur/slugify's rulesets option replaces the defaults. Constructing it with ['rulesets' => ['default', 'german']] broke Cyrillic (empty output) and turned Łódź into od-guide in our test. Use $slugify->activateRuleSet('german') on a default instance instead — that kept everything working.

Ruby results

Tested: Rails' String#parameterize (ActiveSupport 8.1) with default I18n settings.

Input parameterize
Straße & Café strasse-cafe
Привет мир (empty)
Łódź Guide lodz-guide
Don't Stop don-t-stop
Größe Über grosse-uber

parameterize transliterates Latin letters well but, like Django, drops Cyrillic completely unless you add transliteration rules to your I18n locale files.

The big picture

Library Lowercase by default Cyrillic German umlauts (ae/oe/ue) Apostrophes Unique-slug helper
slugify (JS) No — pass lower Yes With locale: 'de' Yes (strict) No
@sindresorhus/slugify Yes Yes Yes Yes Yes
python-slugify Yes Yes Via replacements Splits No
Django slugify Yes Empty No (deletes ß) Yes No
Laravel Str::slug Yes Yes With 'de' Yes No
cocur/slugify Yes Yes Yes Splits No
Symfony AsciiSlugger No — ->lower() Yes With locale Splits No
Rails parameterize Yes Empty No Splits No

Which one should you use?

  • JavaScript/TypeScript (ESM): @sindresorhus/slugify. Best defaults, and the counter helper solves duplicate slugs.
  • JavaScript (CommonJS) or tiny bundles: slugify with { lower: true, strict: true }.
  • Python: python-slugify. Only use Django's built-in if your titles are guaranteed ASCII, or pass allow_unicode=True and accept native-script slugs.
  • PHP / Laravel: Str::slug($title, '-', $locale) — it is already in your project. For non-Laravel apps, cocur/slugify.
  • Ruby on Rails: parameterize for Latin-script content; add I18n transliteration rules (or a gem) for Cyrillic.

Whatever you choose, test it with your own worst-case titles before trusting it with your URLs. The fastest way to see what a correct slug should look like is to paste the same titles into the Slugme.io generator — it handles every input above, including German-style umlauts and native Arabic — and compare. For a whole list at once, use the bulk slug generator.

Want to build your own instead? See our guides for JavaScript, Python and PHP & Laravel.

Methodology: all outputs were produced on October 25, 2026 with the library versions listed above, default options unless stated, and copied verbatim. Some rows (like "Hello World!", where every library returned hello-world) are omitted from the per-language tables for brevity.

Frequently asked questions

What is the best slugify library for JavaScript?

For ESM projects, @sindresorhus/slugify has the best defaults: lowercase output, German-style umlauts, "&" → "and", and a slugifyWithCounter() helper for unique slugs. For CommonJS or the smallest footprint, slugify (by simov) is excellent — just pass { lower: true, strict: true }.

Why does Django's slugify remove some letters?

With allow_unicode=False (the default), Django normalizes to NFKD and drops anything that is not ASCII. Letters that don't decompose into a base letter plus an accent — ß, ł, ø — are deleted, and whole scripts like Cyrillic or Arabic disappear. Use python-slugify if you need transliteration.

Do slugify libraries transliterate Chinese?

Some do. In our test, python-slugify, Symfony's AsciiSlugger and the transliteration npm package converted 影師嗎 to "ying-shi-ma" (pinyin), while slugify, @sindresorhus/slugify, Laravel and cocur returned an empty string.

Should I write my own slugify function?

For a single-language site with simple titles, a 10-line function is fine. As soon as you handle user-generated titles or multiple languages, a maintained library saves you from edge cases — or generate slugs with a dedicated tool and store them.

Keep reading