Urdu language support

[sl49]

“Please improve Firefox’s support for Urdu, including UI translation, spellcheck, and full-page translation from and into Urdu.”

[Jon]

Thanks for submitting an idea to the Mozilla Connect community! Your idea is now open to votes (aka kudos) and comments.

[BuismanSpectra]

“Please improve Firefox’s support for Urdu, including UI translation, spellcheck, and full-page translation from and in

[tedd4x]

Add Urdu to Firefox's built-in translation feature

Firefox Translations has quietly become one of my favourite things about Firefox — local, private, no sending your reading habits to a third-party server. I genuinely appreciate the work that's gone into it.

But I'm puzzled by something: the language list includes Catalan (~10 million speakers), Icelandic (~370,000 speakers), all perfectly valid inclusions, no complaints there. Yet Urdu, spoken by over 500 million people as either a first or second language, is absent. It is the national language of Pakistan (population ~240 million) and one of the 22 scheduled languages of India. Hundreds of millions more across the Middle East, the UK, the US, and Canada read and write it daily.
This isn't a niche ask. Urdu-language news sites, literature archives, government portals, and social content are enormous in volume. Everyday users — researchers, journalists, diaspora communities maintaining ties back home, or just someone trying to read an Urdu newspaper article — currently get nothing from Firefox when they encounter Urdu content, and have to fall back on cloud-based services to fill the gap. That directly contradicts what makes Firefox Translations special.

I understand Urdu uses a right-to-left Nastaliq/Naskh script, which adds some rendering complexity — but Firefox already handles Arabic (another RTL language in the same script family), so the foundational infrastructure should largely be there. Training data shouldn't be a blocker either; Urdu is extraordinarily well-represented on the web and in public corpora. The OPUS corpus, CC-100, and others have robust Urdu datasets. Mozilla's own Common Voice project has Urdu contributions.
I'm not asking for perfection out of the gate. Even an initial model with rough edges would be more useful than the current situation, which is nothing at all.

If there's a technical reason or a prioritization framework that explains the gap, I'd genuinely love to understand it. But if it's simply that no one has pushed for it loudly enough — consider this that push.

[Jon]

(Note: similar ideas have been merged into this thread)