Gnani.ai launches AI voice model that can clone voices in 12 Indian languages

Ganesh Gopalan and Ananth Nagaraj, founders, Gnani.ai

Bangalore-based conversational AI startup Gnani.ai, on February 18, announced the launch of Vachana TTS, a text-to-speech model that can generate human-like speech and clone voices across 12 Indian languages.

The model marks the company’s second release under Inya VoiceOS as part of the India AI Mission.

Launched at the India AI Impact Summit in Delhi, the company said Vachana TTS delivers a Mean Opinion Score (MOS) of 4.23 and a character error rate below 0.6 percent, effectively making it suitable for production-scale use.

Vachana TTS supports Hindi, Bengali, Tamil, Telugu, Kannada, Malayalam, Gujarati, Marathi, Punjabi, Odia, Assamese and Indian English.

The company said the model captures natural rhythm, pronunciation and tone suited to each language.

According to Gnani.ai, independent evaluations show that the model performs better on Indian language naturalness and pronunciation accuracy than existing global providers, while offering lower cost per character.

“Independent evaluations confirm Vachana TTS outperforms existing global TTS providers on Indic language naturalness and pronunciation accuracy, at a substantially lower cost per character - making high-quality voice synthesis economically viable for government and enterprise deployments at population scale,” the company said in a release.

Also, read: Sarvam AI launches 30B and 105B models, says 105B outperforms DeepSeek R1 and Gemini

How Vachana works?

Vachana can copy a person’s voice using just a few seconds of audio.

With less than 10 seconds of someone speaking, the system can recreate their voice and make it sound natural. It keeps the person’s voice traits intact, like how high or low they speak, how fast they talk, and their overall tone and style.

The same voice can also speak in different Indian languages. This means a person’s voice can sound consistent whether it is speaking Hindi, Tamil, Bengali, or other regional languages.

This is especially useful for companies and government departments that operate across many states.

Vachana to work at scale

Vachana TTS is designed to work in challenging conditions, functioning well on low internet speeds and can handle many users at the same time.

It supports two main uses: Real-time voice generation for chatbots and conversational systems and secondly, large-scale voice generation for creating audio content in bulk.

All parts of Vachana TTS are built, trained, and run inside India. The voice data and models stay within Indian data centres, helping organisations meet data localisation and privacy requirements.

Speaking at the summit, Ganesh Gopalan, Co-Founder & CEO, Gnani.ai, said, "We are bringing genuine emotion into synthesized speech - warmth, urgency, empathy - delivered with significantly better accuracy than anything built for Indian languages before, and at a price point that makes it accessible to every enterprise and government body in the country."