Expression Styles

How Style Tags Work

Control the emotional tone of synthesized speech by prepending a style tag to your input text.


Prepending Style Tags

To apply an expression style, prepend a style tag enclosed in angle brackets to the beginning of your input text. The model reads the tag and adjusts prosody, pitch, and rhythm to match the requested emotion.

python
# Happy style
config = TTSConfig(model="zero-indic", voice="Sunita")
result = await client.tts.synthesize(
    "<Happy> आज का दिन बहुत अच्छा है!",
    config=config
)

# News reading style
result = await client.tts.synthesize(
    "<News> The Prime Minister announced new economic reforms today.",
    config=config
)

# Narrative / storytelling style
result = await client.tts.synthesize(
    "<Narrative> Once upon a time, in a land far away, there lived a wise old sage.",
    config=config
)

Default Behavior

Neutral by default:
If no style tag is provided, the model uses the Neutral style automatically. This produces clear, even-toned speech suitable for most use cases.