Bengaluru-based AI startup Sarvam AI has launched a host of both enterprise-based and open-source products as a part of its full-stack Generative AI platform including Sarvam Agents, Sarvam Models and an API for developers. The applications have been trained to support 10 Indic languages including Hindi, Tamil, Telugu, Malayalam, Punjabi, Odia, Gujarati, Marathi, Kannada and Bengali.

Sarvam Agents will be offering multilingual, voice enabled agents who will be available on WhatsApp, over calls or in-app at ₹ 1 per minute for enterprises across sectors like BFSI, healthcare, legal tech and other enterprises.

The startup also released an open-source 2B Indic large language model called Sarvam 2B which has been trained from scratch on an internal dataset of 4 trillion tokens. 

The third product launched was Shuka 1.0, an open-source AudioLM, an audio extension on Meta AI’s Llama 8B model to support Indic languages. 

(For top technology news of the day, subscribe to our tech newsletter Today’s Cache)

It has also launched Sarvam Model for enterprises and developers which will be available via APIs for tasks like translation, speech recognition, speech synthesis and document retrieval. 

Another GenAI tool called A1 was launched for legal firms and lawyers to help with regulatory chat, document drafting, redaction and data extraction. 

“Our mission is to democratise AI and make it accessible to every Indian, regardless of their linguistic and socio-economic background. This platform is a testament to our commitment to bridging the digital divide and fostering innovation in India’s AI landscape, said Pratyush Kumar, co-founder of the startup.

Sarvam has partnered with Hiranandani-backed Yotta Data Services to train its 2B model. 

The startup has raised $41 million in its Series A funding round led by Lightspeed Ventures with participation from Peak XV Partners and Khosla Ventures in December 2023.