Speech To Text Quickstart Guide

Sarvam AI offers two powerful speech models:

Advanced speech recognition model with superior multi-speaker handling and automatic code-mixing support for Indian languages.

Saaras

Domain-aware translation model with enhanced telephony support and intelligent entity preservation.

View our pricing page for detailed information about model-specific pricing and usage tiers.

Supported Audio Formats & MIME Types

For STT & STTT APIs - supports 10+ major audio formats and MIME variants.

Below is a complete list of supported formats and MIME types:

Format Group	Supported MIME Types
MP3 Variants	`audio/mpeg`, `audio/mp3`, `audio/mpeg3`, `audio/x-mpeg-3`, `audio/x-mp3`
WAV Variants	`audio/wav`, `audio/x-wav`, `audio/wave`
AAC Variants	`audio/aac`, `audio/x-aac`
AIFF Variants	`audio/aiff`, `audio/x-aiff`
OGG / Opus Formats	`audio/ogg`, `audio/opus`
FLAC Variants (Lossless)	`audio/flac`, `audio/x-flac`
MP4 / M4A Audio	`audio/mp4`, `audio/x-m4a`
AMR (Narrowband)	`audio/amr`
WMA (Windows Media Audio)	`audio/x-ms-wma`
WEBM (Audio & Video)	`audio/webm`, `video/webm`

Saarika: Our Speech to Text Transcription Model

Saarika is a speech-to-text transcription model that excels in handling multi-speaker content, mixed language content, and conference recordings. It offers automatic code-mixing and enhanced multilingual support, making it ideal for a wide range of applications.

Speech to Text Features

Basic Transcription

With Diarization

Basic Speech to Text Transcription

Convert speech to text with high accuracy. Supports multiple Indian languages and accents. Features include:

Multi-language support
Automatic language detection
High-quality noise filtering
Support for various audio formats

Python

JavaScript

cURL

1 from sarvamai import SarvamAI
2 
3 client = SarvamAI(
4     api_subscription_key="YOUR_SARVAM_API_KEY",
5 )
6 
7 response = client.speech_to_text.transcribe(
8     file=open("audio.wav", "rb"),
9     model="saarika:v2.5",
10     language_code="gu-IN"
11 )
12 
13 print(response)

Check out our detailed API Reference to explore Speech To Text Transcription and all available options.

Saaras Model: Our SOTA Speech to Text Translation Model

Saaras is a domain-aware translation model with enhanced telephony support and intelligent entity preservation. It is designed to handle complex language variations and domain-specific content, making it ideal for call center and telephony applications.

Translation Features

Basic Translation

With Diarization

Speech to Text Translation

Translate speech from any supported Indian language directly into English. Ideal for content localization and international communication. Features include:

Support for major Indian languages
High-quality translations
Preservation of context and tone
Real-time translation capability

Python

JavaScript

cURL

1 from sarvamai import SarvamAI
2 
3 client = SarvamAI(
4     api_subscription_key="YOUR_API_SUBSCRIPTION_KEY",
5 )
6 
7 response = client.speech_to_text.translate(
8     file=open("audio.wav", "rb"),
9     model="saaras:v2.5"
10 )
11 
12 print(response)

Check out our detailed API Reference to explore Speech To Text Translation and all available options.