BHASHINI and IISc Bengaluru release text-to-speech models for Sourashtra

16 October 2026

The Digital India BHASHINI Division, under the Digital India Corporation, has developed text-to-speech models for Sourashtra with the Indian Institute of Science, Bengaluru.
Sourashtra is a minority Indo-Aryan language spoken mainly by the Sourashtra community in Tamil Nadu, and the 2011 Census counted about 2.48 lakh speakers.
The models use the VITS architecture and offer a male and a female voice.
Each voice works with text in both Tamil script and Sourashtra script, making four models in all.
Training used a speech corpus from the SPIRE Lab of the Indian Institute of Science with more than 28 hours of gender-balanced recordings from one male and one female voice artist; the male model was trained on about 10,000 utterances and the female model on about 11,400.
The models generate audio at 22,050 Hz and work directly on character-level text, and the SPICOR TTS 3.0 Sourashtra Corpus is available on the AIKOSH platform.