Sold by: Sarvam AI
Deployed on AWS
Transform audio into text in real-time with our REST API and bidirectional (HTTP/2) streaming API. Built for applications requiring immediate speech processing with minimal delay.
Overview
Speech to Text REST (Saaras V3) - Process short audio files with immediate response. Best for quick transcriptions and testing with a maximum duration of 30 seconds.
Speech to Text Websocket (Saaras V3) - Transform audio into text in real-time with our WebSocket-based streaming API. Built for applications requiring immediate speech processing with minimal delay.
Highlights
- Indic-first ASR across 23 languages - 22 Indian languages + English, with automatic language detection and code-mixed audio support.
- Five output modes in one model - transcribe, translate (to English), verbatim, transliterate (Roman script), and code-mix, selectable per request.
- Production-grade for real workloads - optimized for 8 kHz telephony audio, with utterance-level timestamps and intelligent proper-noun/entity preservation; accepts common audio + telephony codecs. Trained on 1M+ hours of real Indian speech.
Details
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Pricing is based on actual usage, with charges varying according to how much you consume. Subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimension | Description | Cost/host/hour |
|---|---|---|
ml.g6.4xlarge Inference (Batch) Recommended | Model inference on the ml.g6.4xlarge instance type, batch mode | $2.00 |
ml.g6e.4xlarge Inference (Real-Time) Recommended | Model inference on the ml.g6e.4xlarge instance type, real-time mode | $4.00 |
ml.g6.xlarge Inference (Batch) | Model inference on the ml.g6.xlarge instance type, batch mode | $2.00 |
ml.g6e.xlarge Inference (Real-Time) | Model inference on the ml.g6e.xlarge instance type, real-time mode | $4.00 |
ml.g6.2xlarge Inference (Batch) | Model inference on the ml.g6.2xlarge instance type, batch mode | $2.00 |
ml.g6.8xlarge Inference (Batch) | Model inference on the ml.g6.8xlarge instance type, batch mode | $2.00 |
ml.g6.12xlarge Inference (Batch) | Model inference on the ml.g6.12xlarge instance type, batch mode | $8.00 |
ml.g6.16xlarge Inference (Batch) | Model inference on the ml.g6.16xlarge instance type, batch mode | $2.00 |
ml.g6.24xlarge Inference (Batch) | Model inference on the ml.g6.24xlarge instance type, batch mode | $8.00 |
ml.g6.48xlarge Inference (Batch) | Model inference on the ml.g6.48xlarge instance type, batch mode | $16.00 |
Dimensions summary
You pay by the hour for the compute instance that runs the Speech To Text model. Pricing splits into two processing modes. Batch mode runs on ml.g6 instances for transcribing recorded files. Real-time mode runs on ml.g6 and ml.g6e instances for live streaming. Within each mode, you choose an instance size from xlarge up to 48xlarge. Costs scale with the size you pick, since larger instances hold more compute and handle heavier workloads. You are billed per host hour of the instance you select.
Top-of-mind questions for buyers
Real-time mode transcribes live audio streams as speakers talk, useful for voice agents and live captions. Batch mode processes recorded files after the fact, with speaker identification and word-level timestamps. You pick the mode that matches your workload, then choose an instance size within it.
You pay per host hour the instance runs. A fully stopped instance does not accrue software charges. Charges meter running time only, so shutting down instances between jobs stops the hourly software cost. Underlying AWS storage or resource fees may still apply separately.
No. The model, its 22 supported Indian languages, code-mixing, and diarization features are the same across all instance sizes. Instance size affects compute capacity and throughput, not transcription quality. You choose a size to match workload volume and speed needs, not accuracy.
www.sarvam.ai
Helpful?
Vendor refund policy
Contact support@sarvam.ai fot details!
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Amazon SageMaker model
An Amazon SageMaker model package is a pre-trained machine learning model ready to use without additional training. Use the model package to create a model on Amazon SageMaker for real-time inference or batch processing. Amazon SageMaker is a fully managed platform for building, training, and deploying machine learning models at scale.
Deploy the model on Amazon SageMaker AI using the following options:
Version release notes
Launching Sarvam's flagship Speech-To-Text model Saaras:v3.1 supported on L4 and L40s GPU series.
Additional details
Inputs
- Summary
Audio to be transcribed — a call recording, voice note, meeting clip, or broadcast segment. Provide a single audio file in any supported format. Optionally specify the language code (or leave unset for auto-detection) and select an output mode.
Input MIME type multipart/form-data, application/json
- Limitations for input type
- Audio length is capped at 30s for realtime API. For longer audios, websocket api can be used
- Input MIME type
- multipart/form-data, application/json
https://github.com/sarvamai/amazon-sagemaker-examples/tree/main/speech-to-text/saaras-v3.1/data/input
https://github.com/sarvamai/amazon-sagemaker-examples/tree/main/speech-to-text/saaras-v3.1/data/input
Resources
Vendor resources
Support
Vendor support
Saaras (STT) model docs:
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
Deepgram is the enterprise Voice AI platform for building and scaling real time voice applications on AWS. This product listing contains multiple versions of the nova-3 model which can each transcribe a set of languages. See version details for more information.
You will be billed $0.0077/min as described by https://deepgram.com/pricing. Private pricing available upon request.
Our APIs for Nova Speech to Text (STT) are natively available in the new SageMaker Bi-Directional Streaming API. Additional native touchpoints with Amazon Bedrock, Lex, and Amazon Connect make it simple to compose full voice experiences with the cloud services your teams already trust.
Deepgram is the enterprise Voice AI platform for building and scaling real time voice applications on AWS. This product listing contains multiple versions of the nova-3 model which can each transcribe a set of languages. See version details for more information.
You will be billed $0.0092/min as described by https://deepgram.com/pricing. Private pricing available upon request.
Our APIs for Nova Speech to Text (STT) are natively available in the new SageMaker Bi-Directional Streaming API. Additional native touchpoints with Amazon Bedrock, Lex, and Amazon Connect make it simple to compose full voice experiences with the cloud services your teams already trust.
Deepgram is the enterprise Voice AI platform for building and scaling real time voice applications on AWS. This product listing contains multiple versions of the flux model which can each transcribe a set of languages. See version details for more information.
You will be billed $0.0078/min as described by https://deepgram.com/pricing.
Our APIs for Nova Speech to Text (STT) are natively available in the new SageMaker Bi-Directional Streaming API. Additional native touchpoints with Amazon Bedrock, Lex, and Amazon Connect make it simple to compose full voice experiences with the cloud services your teams already trust.
Deepgram is the enterprise Voice AI platform for building and scaling real time voice applications on AWS. This product listing contains multiple versions of the flux model which can each transcribe a set of languages. See version details for more information.
You will be billed $0.0077/min as described by https://deepgram.com/pricing. Private pricing available upon request.
Our APIs for Nova Speech to Text (STT) are natively available in the new SageMaker Bi-Directional Streaming API. Additional native touchpoints with Amazon Bedrock, Lex, and Amazon Connect make it simple to compose full voice experiences with the cloud services your teams already trust.
Deepgram is the enterprise Voice AI platform for building and scaling real time voice applications on AWS. This product listing contains multiple versions of the nova-3 model which can each transcribe a set of languages. See version details for more information. Deepgram charges are billed per request as described by https://deepgram.com/pricing
Our APIs for Nova Speech to Text (STT) are natively available in the new SageMaker Bi-Directional Streaming API. Additional native touchpoints with Amazon Bedrock, Lex, and Amazon Connect make it simple to compose full voice experiences with the cloud services your teams already trust.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.