Why in News?
Researchers at IISc’s Speech, Information and Recognition Engineering (SPIRE) Lab, in collaboration with ARTPARK and Google, have developed SraVaani, described as the first multilingual Indian speech-recognition model trained across 65 Indian languages and dialects.
The model aims to extend Artificial Intelligence (AI)-based automatic speech recognition (ASR) beyond widely supported scheduled languages to several regional and non-scheduled Indian languages.
Key Points
SraVaani has been trained on 65 Indian languages and dialects, including more than 40 languages that are not officially supported by many existing speech-recognition systems.
The model covers 20 scheduled languages and 45 regional languages and dialects, potentially extending speech-AI capabilities to around 25 crore people, based on language data from the 2011 Census, whose languages are inadequately supported by current speech technologies.
The model is designed to address the problem of linguistic under-representation in AI systems, particularly for regional and non-scheduled Indian languages.
SraVaani is freely and publicly available on Hugging Face under an MIT licence. Along with the model, the developers have made a demonstration interface and fine-tuning code available, allowing researchers and developers to experiment with the technology and adapt it for different applications.
CLICK HERE: UPSC PRELIMS & MAINS CURRENT AFFAIRS BASED WEEKLY TEST (ENG & HINDI)
CLICK HERE FOR GS TIMES PRAHAR UPSC GS MAINS TEST SERIES
CLICK HERE: DAILY CURRENT AFFAIRS QUIZ (FOR ALL EXAMINATIONS)
![](https://www.gstimes.in/wp-content/uploads/2026/07/WEEKLY-TEST-RECT.jpg")
