CIS Languages, Codes, and Currencies: A Reference for Website Localization

When setting up hreflang for Kazakhstan, the first instinct is to enter KZ. This is the country code according to ISO 3166-1. The language code is from a different standard, ISO 639-1, and for Kazakh it is kk. For Armenian it is hy, for Georgian it is ka, for Tajik it is tg. None of them match the country code.

The confusion arises because the codes are similar: KZ and kk, AM and hy, GE and ka. This is the main pitfall when setting up hreflang, language switchers, and passing locale to payment systems.

“Kazakh” or “Kazakh” in the menu looks as strange as if “Russian” was written instead of “Русский” in an English interface.

Editorial websites almost always use full native names. tengrinews.kz — «Қазақша / Русский / English», khovar.tj — «Тоҷикӣ / Русский / English», armenpress.am — «Հայերեն» next to a globe icon. Short codes (2–3 characters) are only justified where space is limited: control panels, mobile menus, compact SaaS interfaces.

Kazakh. A developer without Kazakh types K instead of Қ in the code — simply because Қ is not on the Russian keyboard. In the switcher, "Казакша" appears instead of "Қазақша". It looks like a typo, but technically it's a different string: site search will break, sorting will go awry. The abbreviation "КЗ" is incorrect for the same reason; the correct one is "ҚЗ".

«Қазақша» means «in Kazakh» — a colloquial, familiar form. On official websites, «Қазақ тілі» (Kazakh language) is sometimes written, but in the menu, «Қазақша» is shorter. Compare: «Казак» (Cossack, a person) and «Қазақ» (Kazakh people) — the second word simply cannot be written in Russian letters.

Kyrgyz. There is no single standard here, even on official websites. kabar.kg writes «Кыр», other government resources — «Кырг». Both options are correct. Practical advice: choose one and stick to it throughout the interface. «Кырг» is slightly more unambiguous if your menu has several abbreviations starting with «К» (Kazakh, Kyrgyz, Russian).

Both "Kırgızça" and "Kırgız tili" are correct: the first is more colloquial ("to speak Kyrgyz"), the second is more official ("Kyrgyz language"). For a switch, "Kırgızça" is more familiar.

Armenian. This is the only language on the list where the choice of the native name depends on where the audience lives — because Armenian exists in two variants with different orthographies (Wikipedia):

The difference is in one letter: «ե» for Eastern and «է» for Western. But the vocabulary also differs: «airplane» in Eastern is «ինքնաթիռ», in Western — «օդանաւ». For a website in Armenia — Հայերեն. For the Armenian diaspora — Հայերէն. Both have the same ISO code — hy, but ISO 639-3 distinguishes: hye (Eastern) and hyw (Western). The abbreviation «ՀԱՅ» — from Armenian «Հայ» (Armenian): «ՀԱ» without the third letter is not read as a word.

Tajik. The problem is the same as with Kazakh, but immediately visible: Ҷ is in the name of the language itself. If the developer types «ТЧ» instead of «ТҶ» — it's an error right in the switch. Ҷ is not «Ч»: the sound is closer to the English «j» in jump, something between «дж» and a soft «ж». The Ministry of Foreign Affairs of Tajikistan on mfa.tj writes exactly «Ҷ». The word «Тоҷикӣ» also has Ӣ (long «и»), which is also not on the Russian keyboard layout.

Georgian. Developers often try to shorten "ქართული" to "Geo" or "GE" — simply because they don't know how to work with non-Latin fonts in the interface or fear rendering issues. Georgians instantly recognize this: "GE" is perceived as foreign technical code, not as a native word. Mkhedruli is a 1600-year-old script, created in isolation. "გამარჯობა" means "hello." Write "ქართული" in full, even in compact interfaces.

A locale is a "language + country" combination. It defines the format of numbers and dates, the currency symbol, and the sorting order. The locale uses the language code, not the country code: kk_KZ, not kz_KZ. hy_AM, not am_AM. The same logic as with ISO 639-1.

Currency codes are a separate standard, ISO 4217. A few nuances:

Uzbek so'm — official abbreviation, written in lowercase with an apostrophe. In interfaces, "sum" or "UZS" are often used — both options are accepted.

Kazakhstan and Russia share the phone code +7. If the registration form automatically determines the country by phone number, this will break: +7 701 can be both Moscow and Almaty. Request the country as a separate field or determine it by IP.

Georgia has two official languages: Georgian and Abkhazian. This is almost never needed for a commercial website, but if the project is state-owned or oriented towards Abkhazia, take it into account.

If you connect Google Font only with the Cyrillic range, Armenian and Georgian text will not be displayed in the desired font — the browser will substitute a system fallback, and the page will look inconsistent. The Armenian Unicode block is U+0530–U+058F, Georgian is U+10A0–U+10FF. Noto Sans from Google covers both — connect it as a fallback font for these languages.

Most CIS languages use Cyrillic, but three countries switched to Latin after the collapse of the USSR: Uzbekistan in 1993, Azerbaijan in 1991. Cyrillic can still be found there in old documents and among the older generation, but new content is in Latin. Kazakhstan began the transition in 2017, with the new orthography approved in 2023. This is already the third alphabet change in a hundred years: Arabic script, then Janalif (Soviet Latin in the 1930s), then Cyrillic, now Latin again.

Uzbek does not need a separate font — it uses standard Latin plus an apostrophe in O'zbekcha. Most Latin fonts cover this. Just check the symbol ʻ (U+02BB) — it is not everywhere.

Armenian and Georgian are completely independent systems, unrelated to Cyrillic, Latin, or each other. The Armenian alphabet was created in 405 by the monk Mesrop Mashtots for translating the Bible. Georgian Mkhedruli is roughly the same age.

Kazakh and Tajik contain letters that are easily confused with Russian ones on screen, but technically they are different Unicode characters. A site search will not find the word, sorting will be off, and native speakers will notice immediately.

Kazakh has 42 letters — 33 standard plus 9 special: Ә, Ғ, Қ, Ң, Ө, Ұ, Ү, Һ, І. Tajik has its own 6 additional: Ғ, Ҳ, Қ, Ҷ, Ӣ, Ӯ.

Қ and К differ upon closer inspection (Қ has a tail at the bottom) — but Қ is simply not on a Russian keyboard. A person types “Kazaksha” instead of “Qazaqsha” not because they are inattentive, but because the necessary key is unavailable. Or they copy from a source where such a substitution has already been made and do not notice.

Letters in language names and abbreviations are practically important: Қ in “Қазақша” and “ҚЗ”, Ҷ and Ӣ in “Тоҷикӣ” and “ТҶ”, І in Kazakh words. If you insert text from an external source — check it. Insert a suspicious symbol on unicode.org and you will immediately see what that character is.

For most commercial websites, Russian is safer as a default, especially if the audience is over 35 or it's B2B. According to the 2021 census, about 80% of the population speaks Kazakh, and about 70% speaks Russian, and they overlap significantly. For a youth or government audience — Kazakh.

Only if you are specifically working with the Armenian diaspora in Lebanon, France, or the USA. For a project aimed at Armenia, Eastern Armenian (hye) is sufficient. The general code hy is technically accepted for both variants.

hreflang="hy" — for Eastern Armenian (Armenia). For the Western Armenian diaspora, theoretically hyw, but this is a non-standard tag, and not all crawlers process it. It's simpler to have one version in hy, oriented towards Armenia.

Cyrillic. Official document management and most websites are still in Cyrillic. The new Latin orthography was approved in 2023, but there is no mass transition yet. gov.kz monitors the situation.

These are two independent standards. ISO 3166-1 encodes political entities (countries), ISO 639-1 — linguistic (languages). Armenian received the code hy from the Latinized "Hayastan" — the self-designation of Armenia. Georgian — ka from "Kartuli". In some places, the codes match (ru/RU, uz/UZ), in others they don't (hy/AM, ka/GE, kk/KZ). Coincidence is an accident, not a rule.