Protect regional identifiers and local-language context without forcing every startup to build separate regex/NER stacks per market.
Native-script/context rules plus multilingual NER support language-aware detection.
The supplied taxonomy explicitly marks 52 country profiles as strong-format.
The remaining 101 country profiles provide broader contextual coverage while exact local-ID readiness is surfaced in metadata.
The engine can activate multiple language packs for mixed-language text, including English plus local scripts.
NFKC normalization, native digit normalization and script canonicalization make recognition more resilient.
Country metadata can activate source language sections so local identifiers and surrounding language context work together.