BCP 47 Language Tag Parser & Matcher

Parse any language tag — zh-Hant-HK, en-GB-oxendict, de-CH-1901-x-phonebk, i-navajo — into the parts RFC 5646 defines (language, extlang, script, region, variants, extensions, private-use). Validates per-subtag shape, canonicalizes deprecated codes (iw→he, zh-yue→yue), emits the RFC 4647 fallback chain, and runs Basic / Extended Filtering and Lookup against an Accept-Language header so you can debug why your locale switcher picked the wrong one.

Parse a tag

tag

try: en-US zh-Hant-HK de-CH-1901 es-419 sr-Cyrl-RS i-klingon (grandfathered) iw-IL (deprecated) ar-arb-AE (extlang) extensions + private en-GB-oxendict x-my-locale (private-use only)

subtags in order

Canonicalized & normalized

Case per RFC 5646 §2.1.1: language lowercase, script Title-Case, region UPPERCASE, variants and extensions lowercase. Case does not affect equality — matchers must be case-insensitive.

RFC 4647 fallback chain

The Lookup algorithm removes subtags from the right until a match is found. Single-character extension singletons and their subtags are removed as a group.

Match against Accept-Language

client header (or ranges, one per line)

server-supported tags (one per line, top = preferred on tie)

algorithm (RFC 4647)

default when nothing matches

Bulk validate

one tag per line — lines starting with # ignored

Notes

Grandfathered tags (i-navajo, i-klingon, zh-min-nan, …) are legal but should be replaced with their preferred value from the IANA registry when authoring new content.
Extended-language subtags (zh-cmn, ar-arb) are legal but the current recommendation is to use the primary subtag alone (cmn, arb) when the language is known.
3-digit region codes are UN M.49 (419 = Latin America and the Caribbean, 001 = World). 2-letter codes are ISO 3166-1 alpha-2.
The private-use singleton x ends the tag — everything after it is opaque to the outside world.
This tool validates shape, deprecation, and matches per the specs. It does not download the full IANA registry, so unknown-but-well-shaped language / region subtags aren't flagged as invalid (they're marked as unregistered).