Get the app

Chariot Launches Multilingual TTS Model with Support for 11 Indian Languages

AIM (Analytics India Magazine)
Read full story· 3 min
Chariot, an Indian AI lab founded by ex-Google DeepMind and Meta FAIR researchers, launched Chariot-1-TTS-Multilingual, a text-to-speech model supporting 11 Indian languages plus English, with contextual pronunciation and voice cloning. The model outperforms competitors on benchmarks and supports code-mixing and custom voices for applications like voice agents and helplines.

What happened

Chariot, an AI research lab based in Delhi and founded by former Google DeepMind and Meta FAIR researchers, released a new text-to-speech model called Chariot-1-TTS-Multilingual. The model supports 11 Indian languages and English, offering features like contextual pronunciation, code-mixing, and voice cloning. Chariot claims its model outperforms several established competitors on various benchmarks. The model is being made available for use in applications such as voice agents, customer service, and public helplines.

Key details

  • Chariot-1-TTS-Multilingual supports Hindi, Marathi, Bengali, Tamil, Telugu, Kannada, Malayalam, Gujarati, Punjabi, Odia, and Indian English.
  • The model offers over 50 voices and can generate speech from up to 1,000 characters per request.
  • On an English benchmark, the model achieved 93.78% accuracy and the highest plain-English NISQA score among tested models.

Why it matters

Developers and companies building voice-based applications for Indian markets can now access a model that covers a wide range of local languages and supports custom voices. This enables more accurate and natural-sounding automated speech for customer service, educational tools, and public helplines. Organizations needing multilingual voice solutions have a new option that claims higher accuracy and flexibility compared to existing offerings.

#chariot#chariot-1-tts-multilingual#google-deepmind#meta#multilingual-ai#speech-recognition

More in Machine Learning