Laravel Cyrillic ⇄ Latin

Uzbek Cyrillic ⇄ Latin transliteration package for Laravel.

• Context-aware rules, not blind character mapping:
  • е → ye at word start / after vowels (Етти → Yetti), e elsewhere
  • ц → s at word start / after consonants (акция → aksiya), ts after vowels (доцент → dotsent)
  • e → э at word start / after vowels (ekologiya → экология, poeziya → поэзия)
  • ъ/ь before е is a separator: объект → obyekt, съезд → syezd, премьера → premyera
  • с + ҳ is kept apart from the sh digraph: Исҳоқ ⇄ Is’hoq
  • digraphs o‘ g‘ ch sh ya yo yu ts in both directions, including word-final tog‘ → тоғ
  • tutuq belgisi: ma’no ⇄ маъно, e’lon ⇄ эълон
  • abbreviations stay ALL-CAPS: ЧИРЧИҚ → CHIRCHIQ (not ChIRChIQ)
• Accepts every common apostrophe variant: ‘ ’ ʻ ʼ '
• URLs, e-mail addresses, HTML tags and entities are never touched
• Built-in exception dictionary with 200+ stems that break the rules: kompyuter → компьютер, aksiya → акция, rayon → район (not раён), ...
• Suffix-aware stems: one entry covers every Uzbek suffix chain — kompyuterlarida → компьютерларида, rollar → роллар (final ь dropped), effectively tens of thousands of word forms
• Word-level dictionary for your own exception words, with automatic case preservation
• Facade, Str macros, and a framework-free core

Requirements

• PHP 8.1+ (the converter core is PHP 8.1 compatible; Laravel 12 itself requires PHP ≥ 8.2)
• Laravel 12 or 13

Installation

composer require nurbekjummayev/laravel-cyrillic-latin

The service provider is auto-discovered.

Publish the config (optional, for custom dictionaries):

php artisan vendor:publish --tag=laravel-cyrillic-latin-config

Usage

Facade

use NurbekJummayev\LaravelCyrillicLatin\Facades\Transliterator;

Transliterator::toLatin('Ассалому алайкум, Ўзбекистон!');
// "Assalomu alaykum, O‘zbekiston!"

Transliterator::toCyrillic("O'zbekiston — buyuk davlat");
// "Ўзбекистон — буюк давлат"

Transliterator::convert('Салом', 'latin');   // "Salom"
Transliterator::swap('Salom');               // "Салом" (auto-detects the script)
Transliterator::hasCyrillic('Салом');        // true

Note: always import the facade with use. The package intentionally does not register a global Transliterator alias because PHP's intl extension already defines a global \Transliterator class.

Str macros

use Illuminate\Support\Str;

Str::toLatin('Тошкент шаҳри');   // "Toshkent shahri"
Str::toCyrillic('Samarqand');    // "Самарқанд"

Dependency injection

use NurbekJummayev\LaravelCyrillicLatin\Transliterator;

public function show(Transliterator $transliterator)
{
    return $transliterator->toLatin($post->title_cyrillic);
}

Without Laravel

The core has no framework dependencies:

use NurbekJummayev\LaravelCyrillicLatin\Transliterator;

$transliterator = new Transliterator();
$transliterator->toLatin('Фарғона');     // "Farg‘ona"
$transliterator->toCyrillic('Chirchiq'); // "Чирчиқ"

HTML and URLs

Markup, links and e-mails are preserved as-is:

Transliterator::toCyrillic('<b>Toshkent</b> haqida: https://example.uz yoki info@misol.uz');
// '<b>Тошкент</b> ҳақида: https://example.uz ёки info@misol.uz'

Dictionary (exception words)

Built-in exceptions

The Cyrillic soft/hard signs (ь/ъ) and the letter ц are not written in Latin script, so the Cyrillic spelling of many loanwords cannot be reconstructed by letter rules (kompyuter would become компютер, aksiya would become аксия). The package ships a curated list of 200+ stems — months, ц-words and common loanwords — in resources/dictionaries (plain PHP arrays, cached by opcache):

Transliterator::toCyrillic('kompyuter'); // "компьютер" (not "компютер")
Transliterator::toCyrillic('aksiya');    // "акция"     (not "аксия")
Transliterator::toCyrillic('konsert');   // "концерт"   (not "консерт")
Transliterator::toCyrillic('rayon');     // "район"     (not "раён")
Transliterator::toCyrillic('mer');       // "мэр"       (not "мер")

Every stem is suffix-aware — the engine recognizes Uzbek suffix chains (lar, i, ning, da, gacha, ...) and converts them by the letter rules while taking the stem from the dictionary. A word-final ь is dropped before a suffix, as Uzbek Cyrillic orthography requires:

Transliterator::toCyrillic('kompyuterlarida'); // "компьютерларида"
Transliterator::toCyrillic('aksiyalarga');     // "акцияларга"
Transliterator::toCyrillic('rollar');          // "роллар" (роль + лар)

Lookalike words are protected: a stem only matches when the remainder is a valid suffix chain, so rolik → ролик (not роль + ik) and seriya → серия (not сэр + iya).

Disable the built-ins with 'use_default_dictionary' => false in the config.

Abbreviations need no special handling — the letter rules already produce the correct result and dots are left untouched:

Transliterator::toLatin('12 янв. 2026'); // "12 yanv. 2026"

Your own exceptions

Define them in config/cyrillic-latin.php (keys are matched case-insensitively; the original word's capitalization is restored automatically). Your entries always override the built-ins:

'dictionaries' => [
    'latin' => [
        'цех' => 'sex',
    ],
    'cyrillic' => [
        'sex' => 'цех',
    ],
],

Transliterator::toLatin('Цех');  // "Sex"
Transliterator::toLatin('ЦЕХ');  // "SEX"

Architecture

src/
├── Transliterator.php                ← public API: toLatin, toCyrillic, convert, swap
├── CyrillicLatinServiceProvider.php  ← spatie/laravel-package-tools provider
├── Facades/Transliterator.php
├── Contracts/
│   └── Converter.php                 ← convert(string $text): string
├── Support/
│   ├── Alphabet.php                  ← single source of truth for letter mappings
│   └── Dictionary.php                ← exception-word lookup with case preservation
└── Converters/
    ├── ScriptConverter.php           ← word walker; skips HTML/URLs/e-mails
    ├── CyrillicToLatinConverter.php
    └── LatinToCyrillicConverter.php

Both directions are derived from one lowercase letter table in Support/Alphabet.php — uppercase forms, ALL-CAPS digraphs and apostrophe variants are generated from it, so the two directions can never drift apart.

Testing

composer test

License

MIT