What is the correct order of subtags in a BCP 47 language tag?

The canonical order is language-Script-REGION-variant-extension-privateuse, for example zh-Hans-CN or de-1996. Language is lowercase, script is Title-case, region is UPPERCASE, and variants/extensions follow.

When should I include a script subtag like Hans or Cyrl?

Include a script subtag only when it adds information the language subtag does not already imply. zh-Hans versus zh-Hant is meaningful, but en-Latn is redundant since English is suppress-script Latin in the registry.

What is the difference between a region and a variant subtag?

A region subtag is an ISO 3166-1 alpha-2 code (UPPERCASE, like US) or a UN M.49 three-digit number (like 419 for Latin America). A variant subtag (5-8 alphanumerics, or 4 starting with a digit, like 1996 or valencia) marks a dialect or orthography.

What do the u, t and x extension singletons mean?

The 'u' singleton carries Unicode locale extensions such as calendar or collation (en-US-u-ca-gregory). The 't' singleton marks transformed/transliterated content. The 'x' singleton starts a private-use sequence with no registered meaning.

What are grandfathered tags?

Grandfathered tags like i-klingon or zh-min-nan were registered before the current syntax and remain valid for compatibility. Most have a preferred modern replacement (tlh, nan) that you should use in new content.

What is the BCP 47 Language Tag Reference?

Reference for IETF BCP 47 language tags: primary language, script, region, variant and extension subtags, their syntax rules, ordering, and worked examples like zh-Hans-CN and en-US-u-ca-gregory. It runs free in your browser on Gera Tools, with nothing uploaded.

BCP 47 Language Tag Reference

Name: BCP 47 Language Tag Reference
Creator: Gera Tools
License: https://creativecommons.org/licenses/by/4.0/

BCP 47 language tag reference

A BCP 47 language tag identifies a human language together with optional script, region, variant and extension information. It is the value used in HTML lang attributes, hreflang, HTTP Accept-Language, and locale APIs. Tags are built by joining registered subtags with hyphens, drawing on ISO 639 (language), ISO 15924 (script), ISO 3166-1 / UN M.49 (region) and the IANA Language Subtag Registry.

How it works

A well-formed tag follows a fixed order, each part optional except the primary language:

language - Script - REGION - variant - extension - privateuse
   en       Latn      US      1996      u-ca-...     x-foo

Casing is conventional but recommended: language subtags are lowercase (en), script subtags are Title-case 4 letters (Latn), region subtags are UPPERCASE 2 letters or 3 digits (GB, 419), and variants are lowercase. A subtag’s class is determined by its shape: 2-3 letters at the front is the language, exactly 4 letters is a script, 2 letters or 3 digits after that is a region, and 5-8 alphanumerics (or 4 starting with a digit) is a variant.

Extensions begin with a single-letter singleton: u for Unicode locale keywords (calendar, numbering, collation), t for transformed content, and x for private use (everything after x- is opaque). Matching uses these subtags hierarchically — en-GB falls back to en when no exact match exists.

Where BCP 47 tags appear in practice

HTML lang attribute — the most common use. <html lang="en-GB"> tells browsers, screen readers, and spell-checkers which language and locale to apply. Setting only the primary language (lang="en") is usually correct; add the region subtag when British vs. American spelling or date formats matter to your audience.

hreflang links — used in <link rel="alternate" hreflang="fr-CH" href="..."> to tell search engines about alternate-language versions of a page. x-default is a special pseudo-tag for the fallback page. Using zh-Hans vs zh-Hant here correctly targets simplified vs traditional Chinese audiences regardless of their country.

HTTP Accept-Language — browsers send a ranked list of BCP 47 tags in this header. Servers use it for content negotiation: Accept-Language: en-GB,en;q=0.9,fr;q=0.8 prefers British English, then any English, then French.

Intl JavaScript APIs — Intl.NumberFormat, Intl.DateTimeFormat, and Intl.Collator all accept BCP 47 tags. The u extension is especially useful here: en-US-u-nu-fullwide selects full-width digit characters, and en-US-u-ca-hebrew requests the Hebrew calendar.

ICU and CLDR — many formatting libraries (Java’s Locale, .NET’s CultureInfo, Python’s babel) ultimately trace back to BCP 47 under the hood.

Common tags and their correct form

Purpose	Tag	Notes
US English	`en-US`	Most common default for English-language apps
British English	`en-GB`	Date format, spelling and vocabulary variant
Simplified Chinese	`zh-Hans`	Script subtag preferred over `zh-CN` for portability
Traditional Chinese	`zh-Hant`	Covers Taiwan, Hong Kong, and overseas communities
Latin American Spanish	`es-419`	UN M.49 region code for Latin America
Brazilian Portuguese	`pt-BR`	Distinct from European Portuguese (`pt-PT`)
Modern Standard Arabic	`ar`	Script (Arabic) is implied; no subtag needed
Serbian (Cyrillic)	`sr-Cyrl`	Serbia uses both scripts; specify when it matters
Serbian (Latin)	`sr-Latn`	Used in Bosnia and some Croatian-influenced contexts
German (1996 orthography)	`de-1996`	Variant subtag for post-reform German spelling

Tips and examples

Do not over-specify. en is usually better than en-Latn-US-x-mine unless each subtag is genuinely needed; redundant scripts hurt content negotiation.
For Chinese, choose script over region for written content: zh-Hans (simplified) and zh-Hant (traditional) are more portable than zh-CN / zh-TW.
Use the UN M.49 numeric region 419 for “Latin America” when no single country applies, e.g. es-419.
Validate tags against the IANA Language Subtag Registry; only registered subtags are well-formed, with x- private use as the escape hatch for unregistered values.