The human languages NADA covers.

NADA is a shared standard for showing code in a human language. This page lists every language we map and how far its coverage has come. 130+ languages are usable today across 135 mapped languages, built from 4,182,470 translated terms.

Coverage grows continuously. This page is a snapshot, not the full dataset.

Three states.

Every language is in one of three states. We publish the state, not a percentage. A single percentage is easy to misread across 135 languages. Exact per-language coverage is available on request [TO CONFIRM: per-language coverage figures / where they are published].

CompleteCovers the words used in real code. Usable end to end. Refined as vocabulary grows.●

In progressPast the early threshold, still filling gaps. Many already cover more than 70 percent.○

PlannedMapped and queued. Waiting for a contributor or council review.—

Coverage, language by language.

A selection from the 135 mapped languages, grouped by script family. Names appear in English and in each language's own script. The full machine-readable register is part of the open language data [TO CONFIRM: link to the published register / dataset].

European

SpanishEspañolComplete●

FrenchFrançaisComplete●

GermanDeutschComplete●

PortuguesePortuguêsComplete●

ItalianItalianoComplete●

PolishPolskiComplete●

UkrainianУкраїнськаComplete●

GreekΕλληνικάIn progress○

South & Southeast Asian

Hindiहिन्दीComplete●

BengaliবাংলাComplete●

IndonesianBahasa IndonesiaComplete●

Tamilதமிழ்In progress○

TeluguతెలుగుIn progress○

ThaiไทยIn progress○

VietnameseTiếng ViệtComplete●

MarathiमराठीPlanned—

East Asian

Chinese (Simplified)简体中文Complete●

Japanese日本語Complete●

Korean한국어Complete●

Chinese (Traditional)繁體中文In progress○

Middle Eastern & African

ArabicالعربيةComplete●

PersianفارسیIn progress○

TurkishTürkçeComplete●

HebrewעבריתIn progress○

SwahiliKiswahiliIn progress○

HausaHarshen HausaPlanned—

AmharicአማርኛPlanned—

YorubaÈdè YorùbáPlanned—

This is a partial list. 130+ languages are usable today; 79 already pass 70 percent coverage. Proprietary or restricted languages are intentionally not listed.

From a single word to a complete language.

Every language starts from English and is built term by term by people who read it natively. Each contribution is reviewed before it enters the standard, which keeps the vocabulary consistent. No single person or company decides what a word should be.

01A term is proposed. A native reader maps a software word to the right word in their language, in the context where it is used in code.

02The AI council reviews it. An independent panel of models checks the proposal for meaning, consistency, and register, and flags anything that reads wrong to a fluent speaker. A human steward then signs off.

03It enters the language data. Accepted terms join the open CC-BY dataset and ship in the next build, counting toward that language's coverage.

04Coverage moves up a state. As the words used in real code fill in, a language moves from Planned to In progress to Complete.

A language belongs to the people who speak it, and so does its place in the register.

Review thresholds and council composition are documented in the governance memo [TO CONFIRM: link to governance / council documentation].

If your language is missing or incomplete, you can help.

You do not need to be a programmer to contribute. You need to read your language well. Completing one language often advances several related languages at the same time. The data stays free and open for everyone.

Get involved →Or email chris@nadalang.org

Language data CC-BY · 135 languages mapped · the register is public.