Korean API

KoreanG2P applies g2pK phonological rules and optional morphology to prepared Korean text. It does not convert Arabic numerals, Latin words, abbreviations, units, or currencies. Those semantics must be prepared before calling the frontend.

from kokorog2p.ko import KoreanG2P

g2p = KoreanG2P(use_dict=False)
print(g2p.phonemize("안녕하세요"))

Install kokorog2p[ko-mecab] only when morphology/POS analysis is required. The core Korean frontend has no Spokenform or semantic-preparation dependency.

The default isolated-word fast path consults provisioned Lexphon ko:lexhint data before falling back to sentence-level g2pK rules. Korean LexHint data is not bundled or downloaded by KokoroG2P:

lexphon data install ko:lexhint
lexphon data verify ko:lexhint