MeitY Rolls Out VoicERA for Next-Gen Multilingual Voice AI

VoiceRA VoicERA is Launched by MeitY to Multilingual Voice AI Access. VoiceERA came out of stealth mode at the second edition of the India AI Impact Summit 2026 in Pragati Maidan, New Delhi, on 18 February, 2026, organized by MeitY as part of the IndiaAI Mission. 

The platform was launched by Amitabh Kant (former CEO of NITI Aayog) and Amitabh Nag (CEO of Digital India BHASHINI Division) with partners such as EkStep Foundation, Centre for Open Source Software (COSS) at IIIT Hyderabad (not Bengaluru as originally reported) and AI4Bharat. 

It was an international AI governance, ethical AI, and sovereign tech stack summit attended by policymakers, researchers, and industry leaders.

Evolution of BHASHINI

In 2022, Digital India BHASHINI Division was created, which constructs the language translation and voice infrastructure of India based on open-source AI models. The previous milestones are the VoicERA beta that takes place on BHASHINI in November 2025 to be tested. 

BHASHINI is now serving more than 22 scheduled Indian languages and dialects, and is able to provide smooth text-to-speech, speech-to-text, and translation using APIs available to developers and governments. VoiceERA signals the transition to full voice AI stack, bridging the gap in full-scale conversational systems, multilingual systems.

What are the Technical Specifications of VoicERA?

VoicERA has Automatic Speech Recognition (ASR), Neural TTS, Large Language Models (LLM), and Retrieval-Augmented Generation (RAG) pipes, all designed to run on low-latency telephony and real-time interactions. 

Its main elements are domain-specific language models, barge-in detection of natural conversations and integration with UAP (Unified API Platform) of BHASHINI. It can be deployed on Kubernetes to scale to the cloud or an air-gapped environment, and assure the sovereignty of data and adherence to the India DPDP Act. 

The ready-made models are made available to developers in the GitHub repositories under the BHASHINI.

Applications and Use Cases

Voice-based and specifically designed to serve the purposes of the government, VoicERA serves as an energy source to interactive voice Response (IVR) with schemes such as PM-KISAN, Ayushman Bharat, and skill development helplines in local languages. 

They have voice-based crop recommendations to farmers, multilingual redressal of grievances amongst citizens, and chatbots to educate students in remote locations. 

E-commerce customer support and banking queries are also used in the private sector and already piloted in such states as Uttar Pradesh and Tamil Nadu.

VoicERA’s Future in India

VoicERA helps India fulfil its objective of AI-based digital public goods, which can stand to 1.4 billion users by linking with the pillars of Digital India such as DigiLocker and UMANG. 

Future capabilities are emotion recognition, exemption of dialect up to 100 and more variants, and federated learning to achieve privacy enhancement. It has created a startup ecosystem through grants under IndiaAI Mission that attempts to export the model as a global public good to emerging economies.