Voices & languages

46 speaker voices, one male and one female per language, spanning 23 Indic languages and English. The voice parameter controls vocal character; the input text controls the language spoken. They're independent.

Cross-lingual synthesis

Any voice can speak any language. A Tamil-native voice can read Hindi text (with a Tamil-inflected accent); an English-native voice can read Bengali. Useful for brand consistency across markets.

The full catalogue

Language	ISO	Male	Female
Assamese	as	`Bimal`	`Anjana`
Bengali	bn	`Arjun`	`Priyanka`
Bodo	brx	`Daimalu`	`Hasina`
Dogri	doi	`Vishal`	`Neelam`
English	en	`Varun`	`Nisha`
Gujarati	gu	`Rakesh`	`Pooja`
Hindi	hi	`Rajesh`	`Sunita`
Kannada	kn	`Kiran`	`Shreya`
Kashmiri	ks	`Farooq`	`Habba`
Konkani	gom	`Mohan`	`Sarita`
Maithili	mai	`Suresh`	`Meera`
Malayalam	ml	`Krishnan`	`Deepa`
Manipuri	mni	`Tomba`	`Ibemhal`
Marathi	mr	`Siddharth`	`Ananya`
Nepali	ne	`Bikash`	`Sapana`
Odia	or	`Bijay`	`Sujata`
Punjabi	pa	`Gurpreet`	`Simran`
Sanskrit	sa	`Vedant`	`Gayatri`
Santali	sat	`Chandu`	`Roshni`
Sindhi	sd	`Amjad`	`Kavita`
Tamil	ta	`Murugan`	`Thangam`
Telugu	te	`Vishnu`	`Lakshmi`
Urdu	ur	`Salman`	`Fatima`

Cross-lingual examples

The voice controls the vocal character; the input text controls the language. Pick a tab to see four pairings.

python

# English-native voice speaking Hindi
config = TTSConfig(model="zero-indic", voice="Varun")
await client.tts.synthesize("नमस्ते, आप कैसे हैं?", config=config)

python

# Tamil-native voice speaking English
config = TTSConfig(model="zero-indic", voice="Murugan")
await client.tts.synthesize("Good morning, how are you?", config=config)

python

# Hindi-native voice speaking Bengali
config = TTSConfig(model="zero-indic", voice="Sunita")
await client.tts.synthesize("আমি ভালো আছি, ধন্যবাদ।", config=config)

python

# Kannada-native voice speaking Tamil
config = TTSConfig(model="zero-indic", voice="Shreya")
await client.tts.synthesize("வணக்கம், நீங்கள் எப்படி இருக்கிறீர்கள்?", config=config)

Voice selection guide

Pick the voice whose native language matches the primary language of your content, it produces the most natural accent and prosody. Only pick a cross-lingual voice when you have a specific brand reason.

Use case	Recommended male	Recommended female
Hindi IVR / telephony	`Rajesh`	`Sunita`
English voice agent	`Varun`	`Nisha`
Tamil content	`Murugan`	`Thangam`
Telugu content	`Vishnu`	`Lakshmi`
Bengali content	`Arjun`	`Priyanka`
Kannada content	`Kiran`	`Shreya`
Marathi content	`Siddharth`	`Ananya`
Malayalam content	`Krishnan`	`Deepa`
Punjabi content	`Gurpreet`	`Simran`
Gujarati content	`Rakesh`	`Pooja`

Language hint for preprocessing

The optional language field in TTSConfig hints the text preprocessor, it decides how to expand numbers, dates, and script-specific rules. It does not restrict what language the model speaks.

python

# Devanagari number expansion, Hindi date formats
config = TTSConfig(model="zero-indic", voice="Rajesh", language="hi")

python

# English number expansion, English date formats
config = TTSConfig(model="zero-indic", voice="Varun", language="en")

python

# Model auto-handles mixed scripts natively
config = TTSConfig(model="zero-indic", voice="Varun")

Need a custom voice?

Two paths:

Voice cloning

Clone a voice from a 3-6 s reference clip. Works across all 23 languages. Self-service.

Custom-trained voice

For brand voices with strict prosody requirements, Shunya can train a dedicated model. Available on Enterprise plans, contact Shunya Labs ↗.