Getting Started

Models

Sarvam AI gives developers a purpose-built stack for building with Indian languages. From speech-to-text, speech translation, and language translation to high-quality text-to-speech, our models are optimized for India’s multilingual landscape and culturally aligned use cases. With simple, production-ready APIs, developers can quickly build and ship reliable India-first AI experiences without worrying about the underlying complexity.

Model Selection Guide

Saarika: Speech to Text
Saaras: Speech to Text Translate
Bulbul: Text to Speech
Mayura: Text Translation
Sarvam-Translate: Text Translation
Sarvam-M: Chat & Reasoning

Language Support Overview

Core Indic Languages

LanguageCodeLanguageCode
Hindihi-INKannadakn-IN
Bengalibn-INMalayalamml-IN
Tamilta-INMarathimr-IN
Telugute-INPunjabipa-IN
Gujaratigu-INOdiaod-IN
Englishen-IN

Sarvam Translate: Complete 22 Indian Language Support

In addition to the core Indic languages above, Sarvam Translate supports all 22 official Indian languages including:

LanguageCodeLanguageCode
Assameseas-INMaithilimai-IN
Bodobrx-INManipurimni-IN
Dogridoi-INNepaline-IN
Kashmiriks-INSanskritsa-IN
Konkanikok-INSantalisat-IN
Sindhisd-INUrduur-IN

Model Language Support Summary

ModelLanguages
Saarika Speech-to-Text Model11 languages support
Saaras Speech-to-Text Translation Model11 languages support
Bulbul Text-to-Speech Model11 languages support
Mayura Translation Model11 languages support
Sarvam Translate Model22 languages support

Use Cases

Build a multilingual voice assistant

  1. Speech Input: Use Saarika to convert user speech to text
  2. Understanding: Process with Sarvam-M for intelligent responses
  3. Speech Output: Convert responses to speech with Bulbul

Perfect for customer service, smart home devices, and accessibility applications.

Learn how to build a voice agent with LiveKit →