NoaLingua for Chrome
Add to Chrome — Free Add to Chrome

Language hub

What NoaLingua does for each language

NoaLingua offers 28 languages. Eighteen of them have enough verified coverage and enough language-specific detail to justify a page of their own. The rest are rows in the coverage matrix — because publishing ten near-identical pages would help nobody.

English English The English Wiktionary pack is 475 MB — by far the largest NoaLingua downloads, and more than three times the next biggest. It is the clearest illustration of what local-first actually costs in disk space. Español Spanish Spanish has the largest UniMorph dataset of any supported language: 1,196,245 rows, of which 1,048,576 are verb forms. Français French French UniMorph is verbs only. Measured at 367,732 rows, every one is V, V.PTCP or V.CVB — there is not a single noun or adjective in it. Deutsch German German UniMorph is one of the few that is genuinely balanced: 519,143 rows split across nouns (219,104), verbs (198,913) and adjectives (87,503) — so case tables for nouns and adjective endings both resolve. Italiano Italian Italian UniMorph is verbs only — 509,574 rows, all V, V.PTCP or V.CVB. Nouns and adjectives come from the Wiktionary pack or not at all. Português Portuguese Portuguese has one of the seven bundled frequency lists. Русский Russian Russian has one of the seven bundled frequency lists — the only Cyrillic-script language that does. 日本語 Japanese Japanese UniMorph is the smallest of all at 1 MB, while the Wiktionary pack is 45 MB — nearly all the useful grammar detail for Japanese comes from the latter. 中文 Chinese There is no Chinese UniMorph dataset, and that is not a gap — Chinese is an isolating language, so there is genuinely nothing to conjugate or decline. العربية Arabic Arabic UniMorph is 7 MB against a 47 MB Wiktionary pack. Türkçe Turkish Turkish UniMorph is 32 MB — among the largest — which is what agglutination looks like as a file size: one stem generates a very large number of forms. Српски Serbian There is no Chrome on-device translation model for English–Serbian. It is unavailable, not merely undownloaded — which is precisely why the extension keeps YouTube’s own caption translation as a fallback. Hrvatski Croatian Croatian is one of two languages the handbook names as covered by YouTube’s own caption translation. Tiếng Việt Vietnamese Vietnamese has neither a UniMorph nor a kaikki pack. The UniMorph absence is honest rather than a gap — Vietnamese is analytic, so there is nothing to conjugate — but the missing dictionary means definitions come from the online sources only. ไทย Thai Thai has neither a UniMorph nor a kaikki pack — the same shape as Vietnamese, and for the same reason on the grammar side: Thai is analytic. Magyar Hungarian Hungarian UniMorph is 37 MB — the second largest of any language — which is what heavy agglutination costs in stored forms. Suomi Finnish Finnish has no grammar pack, and the reason is mundane: UniMorph publishes Finnish across two files instead of one, and the downloader assumes one. It was left out deliberately rather than shipping half the data silently. Norsk Norwegian Norwegian grammar comes from UniMorph’s `nob` dataset — Bokmål, chosen over the much smaller Nynorsk (`nno`). If you are learning Nynorsk, the tables will not match what you read.

The other ten

Languages without their own page

These are fully supported and appear in the coverage matrix with the same verified data as everything else. They do not have a page because we have fewer than three things to say about them that are not already said elsewhere — and padding a page to reach three is exactly how a site turns into a doorway farm.

See the full coverage matrix →