English API

The English frontend phonemizes prepared English text with an externally provisioned Lexphon dictionary and optional backend/model controls. It does not expand numbers, currencies, dates, units, or abbreviations.

from kokorog2p.en import EnglishG2P

g2p = EnglishG2P(language="en-us", lexicons="gold", use_spacy=False)
print(g2p.phonemize("Hello world"))

Install the selected asset before construction:

lexphon data install en-us:gold en-us:lexhint en-gb:gold en-gb:lexhint
lexphon data verify en-us:gold en-us:lexhint en-gb:gold en-gb:lexhint

Use get_g2p("en-us") or get_g2p("en-gb") for factory construction. The frontend preserves supplied text and token offsets; semantic preparation belongs to the caller.

Lexicon controls

English exposes gold and an explicitly selected lexhint source. Gold values use the kokoro-v1 encoding; LexHint values use IPA and pass through the same English realization profile after decoding. Use lexicons=() for fallback-only operation. There is no English silver tier or runtime API for loading packaged dictionaries. use_spacy and explicit local model settings control optional POS-aware tokenization.