Slug Generator

Fund package maintenance!

Installs: 1 503 344

Dependents: 21

Suggesters: 0

Security: 0

Stars: 798

Watchers: 25

Forks: 54

Open Issues: 12

v1.1.1 2020-10-19 15:25 UTC


Build Status Coverage Packagist Version Downloads MIT License

This library provides methods to generate slugs for URLs, filenames or any other target that has a limited character set. It’s based on PHPs Transliterator class which uses the data of the CLDR to transform characters between different scripts (e.g. Cyrillic to Latin) or types (e.g. upper- to lower-case or from special characters to ASCII).


use Ausi\SlugGenerator\SlugGenerator;

$generator = new SlugGenerator;

$generator->generate('Hello Wörld!');  // Output: hello-world
$generator->generate('Καλημέρα');      // Output: kalemera
$generator->generate('фильм');         // Output: film
$generator->generate('富士山');         // Output: fu-shi-shan
$generator->generate('國語');           // Output: guo-yu

// Different valid character set, a specified locale and a delimiter
$generator = new SlugGenerator((new SlugOptions)
$generator->generate('Äpfel und Bäume');  // Aepfel_und_Baeume


To install the library use Composer or download the source files from GitHub.

composer require ausi/slug-generator

Why create another slug library, aren’t there enough already?

There are many code snippets and some good libraries out there that create slugs, but I didn’t find anything that met my requirements. Options are often very limited which makes it hard to customize for different use cases. Some libs carry large rulesets with them that try to convert characters to ASCII, no one uses Unicode’s CLDR which is the standard for transliteration rules and many other transforms.

But most importantly no library was able to do the “correct” conversions, like Ö-Äpfel to OE-Aepfel for German or İNATÇI to inatçı for Turkish. Because the CLDR transliteration rules are context sensitive they know how to correctly convert to OE-Aepfel instead of Oe-Aepfel or OE-AEpfel. CLDR also takes the language into account and knows that the turkish uppercase letter I has the lowercase form ı instead of i.


All options can be set for the generator object itself new SlugGenerator($options) or overwritten when calling generate($text, $options). Options can by passed as array or as SlugOptions object.

delimiter, default "-"

The delimiter can be any string, it is used to separate words. It gets stripped from the beginning and the end of the slug.

$generator->generate('Hello World!');                         // Result: hello-world
$generator->generate('Hello World!', ['delimiter' => '_']);   // Result: hello_world
$generator->generate('Hello World!', ['delimiter' => '%20']); // Result: hello%20world

validChars, default "a-z0-9"

Valid characters that are allowed in the slug. The range syntax is the same as in character classes of regular expressions. For example abc, a-z0-9äöüß or \p{Ll}\-_.

$generator->generate('Hello World!');                             // Result: hello-world
$generator->generate('Hello World!', ['validChars' => 'A-Z']);    // Result: HELLO-WORLD
$generator->generate('Hello World!', ['validChars' => 'A-Za-z']); // Result: Hello-World

ignoreChars, default "\p{Mn}\p{Lm}"

Characters that should be completely removed and not replaced with a delimiter. It uses the same syntax as the validChars option.

$generator->generate("don't remove");                         // Result: don-t-remove
$generator->generate("don't remove", ['ignoreChars' => "'"]); // Result: dont-remove

locale, default ""

The locale that should be used for the Unicode transformations.

$generator->generate('Hello Wörld!');                        // Result: hello-world
$generator->generate('Hello Wörld!', ['locale' => 'de']);    // Result: hello-woerld
$generator->generate('Hello Wörld!', ['locale' => 'en_US']); // Result: hello-world

transforms, default Upper, Lower, Latn, ASCII, Upper, Lower

Internally the slug generator uses Transform Rules to convert invalid characters to valid ones. These rules can be customized by setting the transforms, preTransforms or postTransforms options. Usually setting preTransforms is desired as it applies the custom transforms prior to the default ones.

How Transform Rules (like Lower or ASCII) and rule sets (like a > b; c > d;) work is documented on the ICU website:

$generator->generate('Damn 💩!!');                                           // Result: damn
$generator->generate('Damn 💩!!', ['preTransforms' => ['💩 > Ice-Cream']]);  // Result: damn-ice-cream

$generator->generate('©');                                          // Result: c
$generator->generate('©', ['preTransforms' => ['© > Copyright']]);  // Result: copyright
$generator->generate('©', ['preTransforms' => ['Hex']]);            // Result: u00a9
$generator->generate('©', ['preTransforms' => ['Name']]);           // Result: n-copyright-sign


Thanks to Blackfire for sponsoring performance profiling tools for this project.