Audio Processing

Language Preprocessing

Pass an ISO 639 language code to activate the text preprocessor for a specific language, improving number expansion, abbreviation handling, and transliteration.


How it works

The language parameter is optional. The model handles mixed scripts natively, but providing the language code activates a dedicated text preprocessor that can improve pronunciation accuracy for edge cases like currency symbols, large numbers, and domain-specific abbreviations.

PROPERTYVALUE
Parameterlanguage
Typestring
Defaultnull

Examples

Hindi text

python
config = TTSConfig(
    model="zero-indic",
    voice="Sunita",
    language="hi",   # Hindi preprocessor
)
result = await client.tts.synthesize(
    "आपका ऑर्डर № 12345 की कीमत ₹1,299 है।",
    config=config,
)

English text

python
config = TTSConfig(
    model="zero-indic",
    voice="Nisha",
    language="en",   # English preprocessor
)
result = await client.tts.synthesize(
    "Your balance is $4,200.50 as of Jan 1, 2025.",
    config=config,
)

Mixed text (no language code needed)

python
config = TTSConfig(
    model="zero-indic",
    voice="Sunita",
    # language omitted — model handles mixed scripts natively
)
result = await client.tts.synthesize(
    "Hello! आपका order number 56789 ship हो गया है।",
    config=config,
)