The project also aims to preserve folklore, oral traditions, and cultural heritage using optical character recognition technology.
Developing an AI-based translation solution for a popular language is child’s play for Natural Language Processing experts. But it is not the same case with hinterland languages, which don’t have their own script, and face extinction.
A multi-institute effort, driven by the Union Ministry of Tribal Affairs, has resulted in the AI-based Android app Adi Vaani, a translation platform for Santali, Bhili, Mundari, and Gondi languages. The platform enables real-time translation of both text and speech between Hindi/English and the tribal languages. The app is going to be avaible on Google’s appstore Play Store soon.
The app will also help the respective communities to get access to information on various government schemes that are targetted at them.
Biggest challenge
“The biggest challenge is that Adivasi languages don’t have their own script. They write in the script belonging to the nearest mainland language. There is not much content available too to train the machines,” Radhika Mamidi, a Professor at the IIIT-Hyderabad’s Language Technologies Research Centre, told businessline.
The country is home to 461 tribal languages in 71 distinct tribal mother tongues. Of these, as many as 81 are considered vulnerable and 42 are critically endangered in the absence of content and intergenerational transmission gaps.
“Adi Vaani is an attempt to address this challenge. It ensures AI-based digitisation, preservation and revitalisation of tribal languages. The beta version of this app will soon be available on Google’s app store Play Store. “An iOS application in the works,” she said.
Besides IIIT-Hyderabad, IIT Delhi, BITS Pilani, IIIT Nava Raipur, and Tribal Research Institutes (TRIs) in Jharkhand, Odisha, Madhya Pradesh, Chhattisgarh, and Meghalaya, took part in the project.
With the help of TRIs, the participant institutes recruited volunteers from the respective languages and developed content and voice samples to build the app.
The project also aims to preserve folklore, oral traditions, and cultural heritage using optical character recognition technology.
The researchers also developed a text-to-speech tool for Santali, Mundari, and Bhili languages, while the tool for Gondi is under development.
Published on September 4, 2025