Sold by: Murf Inc
Deployed on AWS
Murf Falcon 2.1 is a low-latency text-to-speech model for voice agents, IVR and narration. It streams the first audio in 44 ms (median, one stream on ml.g6e.xlarge) and speaks 42 locales with 141 voices, many of them in several languages. It runs entirely inside your AWS account with network isolation, so text and audio never leave your VPC.
Overview
Murf Falcon 2.1 turns text into natural, expressive speech fast enough for live conversation. Use it for AI voice agents and contact centers, IVR prompts, in-app voice, e-learning and content narration.
Voices and languages: 141 voices in 42 locales across 35 languages, including English (US, UK, India, Australia, Scotland), Spanish, French, German, Portuguese, Italian, Dutch, Polish, Japanese, Korean, Chinese, Arabic, Hindi and 10 more Indian languages. Many voices speak several languages natively and offer more than one speaking style, so one voice can serve a multilingual product.
Performance, measured in November 2026 on one ml.g6e.xlarge (NVIDIA L40S) with 157-character streaming requests: median time to first audio is 44 ms for one stream, 60 ms for 10 concurrent streams and 84 ms for 32. Audio is generated about 10x faster than real time. Intelligibility: 1.0-1.4% word error rate when English output for ten test passages is transcribed by a speech recognizer, across three voices.
Features: real-time, response streaming, bidirectional streaming (feed LLM tokens as they arrive) and batch transform; word-level timestamps; WAV, MP3, Opus, FLAC, PCM and telephony (mu-law, A-law) audio at 8-48 kHz; speed, pitch and volume control; built-in text normalization for numbers, dates, currencies and abbreviations. Generated audio carries an inaudible AudioSeal watermark that marks it as synthetic speech.
Privacy: the package deploys with SageMaker network isolation. Nothing is downloaded at start-up and no data is sent to Murf or any third party; the endpoint logs only character counts and timings, never text.
Training data: licensed and proprietary studio speech recordings across 35 languages.
Limitations: input is plain text, up to 3,000 characters per real-time request and 20,000 per streaming request. Romanized Hindi is not supported; send Devanagari text. Voices are fixed to the catalog: custom voice cloning is not included.
Pricing: real-time and streaming usage is billed per 1,000 input characters; batch transform is billed per instance-hour. Both are in addition to the SageMaker instance cost.
Highlights
- 44 ms median time to first audio when streaming on ml.g6e.xlarge, and up to 32 concurrent real-time streams per instance.
- 141 voices in 42 locales across 35 languages, with multiple speaking styles; many voices speak several languages natively.
- Runs entirely in your AWS account with network isolation: no text or audio leaves your VPC, and output is watermarked as synthetic 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 |
|---|---|---|
ml.g6e.xlarge Inference (Batch) Recommended | Model inference on the ml.g6e.xlarge instance type, batch mode | $85.68/host/hour |
inference.count.m.i.c Inference Pricing | inference.count.m.i.c Inference Pricing | $0.015/request |
ml.g5.xlarge Inference (Batch) | Model inference on the ml.g5.xlarge instance type, batch mode | $34.80/host/hour |
ml.g6.xlarge Inference (Batch) | Model inference on the ml.g6.xlarge instance type, batch mode | $31.20/host/hour |
Vendor refund policy
Usage charges are non-refundable. For billing questions contact support@murf.ai .
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
Initial release.
Additional details
Inputs
- Summary
A JSON object (Content-Type: application/json) with the text to speak, the voice and optional audio settings.
- Limitations for input type
- 3,000 characters per real-time request, 20,000 per streaming request. One request per batch transform input object.
- Input MIME type
- application/json
https://raw.githubusercontent.com/murf-ai/murf-falcon-sagemaker/main/samples/realtime_request.json
https://raw.githubusercontent.com/murf-ai/murf-falcon-sagemaker/main/samples/batch_input/welcome.json
Input data descriptions
The following table describes supported input data fields for real-time inference and batch transform.
Field name | Description | Constraints | Required |
|---|---|---|---|
text | Text to speak. | Up to 3,000 characters per real-time request, 20,000 when streaming. | Yes |
voice_id | Voice to use, e.g. en-US-natalie. | One of the IDs returned by list_voices. | Yes |
format | Audio format: wav, mp3, ogg_opus, flac, pcm_s16le, pcm_f32le, mulaw or alaw. | flac cannot be streamed. | No |
sample_rate | Output sample rate in Hz. | 8000, 16000, 22050, 24000, 44100 or 48000. | No |
locale | Language for the voice to speak, e.g. hi_IN. | One of the voice's supported_locales. | No |
style | Speaking style, e.g. Conversational. | One of the voice's styles for that locale. | No |
speed | Speaking rate. | 0.5 to 2.0 | No |
pitch | Pitch multiplier. | 0.5 to 2.0 | No |
stream | Stream audio as it is generated (InvokeEndpointWithResponseStream). | - | No |
stream_format | raw audio bytes or ndjson audio chunks with word timestamps. | raw or ndjson | No |
Resources
Support
Vendor support
Email support@murf.ai with your AWS Region, the model package version and the request_id from the response. Getting started guide, sample notebook and troubleshooting:
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
Murf Falcon enables you to build voice agents that are ultra-fast, expressive, scalable, and highly cost-efficient. It has 55 ms model latency and 130 ms end-to-end latency. It supports 35+ languages and pricing starting at 1 cent per minute.
Murf API offers enterprise-grade AI voice generation capabilities with its industry-leading text-to-speech models, transforming multimedia and conversational experiences. Our API achieves over 99% pronunciation accuracy and offers unparalleled speech customization through styles, pauses, duration matching and variation controls.
Speech-to-text for voice agents, contact centers and transcription in English, six European languages and 11 Indian languages. Finch streams transcripts while people speak, delivers each final transcript within about 250 ms of the end of speech, and runs inside your AWS account with network isolation, so audio never leaves your account.
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 aura-2 model which can each speak a set of languages and voices. 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.