Bhashini to power on-device multilingual capabilities on Intel AI PCs

Intel and BHASHINI bring on-device multilingual translation and transcription to AI PCs, with Vidyalekha running offline on Core Ultra-powered laptops for education use cases

Intel India has partnered with the Digital India BHASHINI Division under the Ministry of Electronics and Information Technology (MeitY) to bring on-device multilingual capabilities to Intel-powered AI PCs. The two organisations announced that BHASHINI’s Vidyalekha tool will now be available on AI PCs powered by Intel Core Ultra series processors, enabling real-time translation and transcription across multiple Indian languages.

 

The integration is aimed primarily at education use cases. According to the companies, Vidyalekha is designed to work offline on laptops and can transcribe English lectures and translate them into a student’s preferred Indian language. The tool is intended to assist college students who may have studied in regional languages and face difficulties following instructions in English.

 

How it works

The collaboration involves optimising BHASHINI and AI4Bharat’s translation and transcription models to run on Intel’s AI PC xPU architecture. This enables AI workloads to execute locally on devices rather than relying on cloud infrastructure.

 

By running these models directly on AI PCs, the system is designed to deliver lower latency performance and operate without network connectivity. Local processing also helps preserve privacy and data confidentiality, as datasets do not need to be transmitted to external servers.

 

The companies demonstrated automatic speech recognition, text-to-speech and translation models running in real time on Intel Core Ultra-powered AI PCs at the India AI Impact Summit 2026 in New Delhi, last month.

What is BHASHINI

BHASHINI (BHASHa INterface for India) is a government-backed language technology platform developed as part of the Digital India initiative. It focuses on building AI tools for translation, speech recognition and text-to-speech across Indian languages, with an emphasis on use in public-facing and government digital services.

 

The platform currently supports more than 36 languages for text-based applications and over 22 languages for voice interactions. It has been integrated into multiple government portals and applications to enable users to access services, complete forms and receive official information in regional languages, rather than being limited to English or a small set of widely used languages.

Last month, India-based cloud provider Yotta Data Services migrated the BHASHINI platform from a global hyperscaler to domestic infrastructure. BHASHINI is now hosted on Yotta’s Government Community Cloud and Shakti Cloud, operating entirely on Indian cloud and GPU infrastructure. The migration, which reportedly took two to three months, covered the platform’s full AI stack — including datasets, models, APIs, containerised services, orchestration pipelines, databases and storage.