AI4Bharat is a research lab at IIT Madras which works on developing open-source datasets, tools, models and applications for Indian languages. Over the past few years, the lab has done cutting-edge work in Indian language technology which has been published in top conferences around the world. Some of the top contributions from the lab are IndicCorp, BPCC, Shrutilipi, Kathbath, IndicBERT, IndicTrans, IndicXlit, IndicWav2Vec, IndicWhishper, and TTS. All our models, tools, datasets and applications are fully open source and can be used by students, academia, industry, governments, startups and other organizations.

The work done at AI4Bharat is closely aligned with the goals of the ambitious Digital India Bhashini Mission (Bhashini) being run under the aegis of the Ministry of Electronics and Information Technology (MeitY). Prof. Mitesh Khapra who heads AI4Bharat is also serving at the PI for the Data Management Unit (DMU) at Bhashini. All datasets listed on this website have been created using a generous grant given by Bhashini (unless explicitly mentioned otherwise). Further, several models listed on the website and mentioned above have been trained using CDAC Pune’s supercomputer, ParamSiddhi.AI4Bharat is also funded by Nilekani Philanthropies and EkStep Foundation for setting up the Nilekani Centre at AI4Bharat. Among other things, the Centre also assists Bhashini in its mission as a Bhashini Sahayogi, by employing researchers, engineers and data collection leads who help in managing the large data collection efforts at DMU. The Centre also helps in developing tools, models and applications which are released in the open source with fully permissible licenses to boost the language technology ecosystem in the country.AI4Bharat has also received generous support from Microsoft, Google and Nvidia in the form of compute credits and/or unrestricted research grants.