Bhashini and IISc Bengaluru Develop AI-Based Text-to-Speech Models for Sourashtra Language

A division of the Digital India BHASHINI program in collaboration with the Indian Institute of Science (IISc), Bengaluru, has designed artificial intelligence-based text-to-speech (TTS) models for Sourashtra, which is an Indo-Aryan language used predominantly by the Sourashtra ethnic group of people residing in Tamil Nadu. This innovation has been announced by the Ministry of Electronics and Information Technology (MeitY) on October 16, 2026.

This is an important development in relation to a language that lacks digitized resources. The TTS models will be able to convert written text in the language into spoken audio.

Four AI Models Support Two Scripts

The newly developed models use the VITS (Variational Inference with adversarial learning for end-to-end Text-to-Speech) architecture, which generates speech directly from written text.

The project includes four models offering male and female voices, with each voice supporting text written in both Tamil script and Sourashtra script. This allows users to access spoken content regardless of which of these two scripts they use to write the language.

The models were trained using a dedicated speech corpus developed at IISc Bengaluru‘s Signal Processing Interpretation and Representation (SPIRE) Lab. The dataset contains more than 28 hours of gender-balanced speech recordings, including composed sentences and material drawn from literature.

The male voice model was trained on approximately 10,000 utterances, while the female voice model used around 11,400 utterances. The resulting models are designed to generate natural-sounding speech directly from text. 

Sourashtra Speech Dataset Available on AIKOSH

The speech dataset, known as the SPICOR TTS 3.0 Sourashtra Corpus, is now available through AIKOSH, BHASHINI‘s platform for language datasets and AI resources.

Making the corpus accessible can help researchers, developers and educational institutions build additional speech technologies for Sourashtra. The resource also provides a foundation for future research into speech synthesis for Indian languages that have limited digital representation.

How the Technology Can Help Sourashtra Speakers

The new models could support several practical applications. Educational platforms may use them to create audio learning materials, while accessibility tools could help users listen to written content instead of relying exclusively on text.

The technology may also assist in documenting literary material and developing digital resources for younger generations of Sourashtra speakers. Researchers can use the available speech data to explore improvements in low-resource language processing.

According to the 2011 Census, approximately 2.48 lakh people spoke Sourashtra. Expanding digital tools for the language could help improve its visibility in an increasingly technology-driven world. 

Strengthening India’s Multilingual AI Ecosystem

The collaboration reflects BHASHINI’s broader objective of making digital services accessible across India’s linguistic diversity. The platform supports speech, translation and other language technologies for Indian and international languages.

Developing AI tools for underrepresented languages is an important step towards ensuring that technological progress benefits communities beyond widely used languages.

The Sourashtra text-to-speech models provide a starting point for further innovation in education, accessibility and language preservation. Their long-term impact will depend on adoption by developers, institutions and the Sourashtra-speaking community.

Defence Breakthroughs, Travel Alerts & Trade Tensions

How Is India-France Co-Development Shaping Future Defence?
Discover how India and France are strengthening their defence partnership through joint development and next-generation military technologies.

What Changes for Indians Using UAE Services?
Check out the latest updates on passport, visa, and attestation services moving to new centres, including what applicants need to know.

How Could US Drug Tariffs Affect India?
Explore how proposed US tariffs on generic drug imports could impact Indian pharmaceutical exports, pricing, and the wider pharma industry.

How Can You Claim Unclaimed Bank Deposits?
Check out the RBI’s guidelines and follow a step-by-step guide to finding and claiming unclaimed money through the UDGAM portal.

How Is Kaziranga Protecting Wildlife During Floods?
Discover the wildlife rescue operations, habitat monitoring measures, and conservation efforts helping protect animals in Kaziranga National Park.