IISc Unveils Open-Source Voice AI Model Supporting 65 Indian Languages

The Indian Institute of Science (IISc) has launched a new open-source voice artificial intelligence model capable of understanding and transcribing speech across 65 Indian languages and dialects, marking a significant step toward improving digital accessibility in one of the world’s most linguistically diverse countries.

The model, named SraVaani 1.0, has been developed by researchers working on speech recognition technologies for Indian languages. It is designed to address a long-standing challenge in the AI sector, where most speech recognition systems offer limited support for India’s vast linguistic landscape and often perform poorly in low-resource and regional languages.

According to researchers, the system has been trained on tens of thousands of hours of speech data collected from across the country. The project draws on the VAANI corpus, a large repository of Indian speech recordings, and combines data from multiple public datasets to improve recognition accuracy across languages and dialects.

One of the key features of the model is its support for several low-resource and tribal languages that currently have little or no representation in mainstream AI systems. Researchers said the technology aims to make voice-based digital services more inclusive by extending support beyond major languages such as Hindi, Bengali, Tamil and Telugu.

The launch comes amid a broader push in India to develop indigenous AI technologies tailored to local needs. Government-backed initiatives and academic institutions have increasingly focused on building language technologies that can bridge the digital divide and make online services accessible to users who are more comfortable communicating in regional languages.

By releasing the technology as open source, IISc hopes to encourage researchers, developers and startups to build new applications on top of the platform. The initiative is expected to strengthen India’s growing ecosystem of homegrown AI solutions while ensuring that advances in voice technology benefit speakers of a wider range of Indian languages.