Amazon Polly Generative Engine Receives Major Upgrade
The Amazon Polly Generative engine has received a major production-ready upgrade, introducing 14 new voices across nine locales, general availability for Hindi and Finnish, and enhanced expression without requiring any code changes.
Radwan AbdulhadiEditor and technology writer
Major Upgrade Brings Enhanced Expression and Accuracy
AWS has announced a major upgrade to the Amazon Polly Generative engine, designed to deliver more natural, expressive, and accurate speech output with low latency. The fully managed service converts text into natural-sounding speech for contact centers, large language model (LLM)-based agentic applications, media and entertainment, e-learning, and accessibility platforms. According to the AWS Machine Learning Blog, the release enhances emotion and prosody based on text input without requiring manual setup, allowing voices to dynamically adjust across neutral, uncertain, and surprised styles.
The updated engine also improves handling for structured elements, ensuring precise speech generation for numbers, currency, dates, times, and addresses. Developers can review specific implementation patterns and technical parameters directly through the official Amazon Polly documentation to optimize their workflows.
New Languages, Voices, and Regional Expansion
The update expands the broader catalog for the Amazon Polly Generative engine, bringing the total to 57 generative voices across 11 languages and 25 locales. A total of 14 new voices have been added across nine locales, including US English, Australian English, Indian English, German, US Spanish, Brazilian Portuguese, Korean, Hindi, and Finnish.
Notably, this release marks the general availability of Hindi and Finnish generative voices for the first time. Furthermore, the engine's regional footprint has expanded to AWS GovCloud (US-West), bringing the total number of supported AWS regions to 11.
Drop-In Production Compatibility
AWS designed the upgrade to be a seamless drop-in replacement for existing implementations. Voice names and API calls remain completely unchanged, ensuring that current enterprise applications continue to function without modification. For existing generative voices, the upgrade happens automatically in place.
New applications can access the expanded voice catalog through the standard SynthesizeSpeech API. Additionally, real-time applications such as voice agents and custom game characters can utilize the Bi-Di Streaming API StartSpeechSynthesisStream to send text and receive synthesized audio simultaneously.
Consistent Enterprise Pricing
Despite the significant quality and performance upgrades, AWS maintains competitive pay-as-you-go pricing for the service. The model update delivers improved voice accuracy and natural expression at the same baseline price, with no cost increases, no new stock-keeping units (SKUs), and no alterations required for current integrations.
Developers can continue caching and replaying static prompts at no extra charge. Complete details regarding pricing structures and cost tiers can be found on the official Amazon Polly pricing page.
Sources
- AWS Machine Learning BlogAmazon Polly delivers a major upgrade to its Generative engine
Continue chronologically
Related entity coverage
Related stories
IskayPet Unifies Corporate Knowledge with Amazon Bedrock
IskayPet built a multilingual corporate knowledge assistant using Amazon Bedrock Managed Knowledge Bases to streamline employee operations across Iberia.
Managing Amazon Bedrock Costs: A FinOps Playbook
Explore an end-to-end operational model for FinOps and platform teams to manage, attribute, and optimize Amazon Bedrock generative AI costs.
Deploy Decision Models with llama.cpp on Amazon SageMaker AI
Learn how to deploy decision models and OpenAI-compatible chat completions with llama.cpp on Amazon SageMaker AI using CPU or GPU infrastructure.
Build Multi-Agent Pipelines with Amazon Bedrock AgentCore
Learn how to build a multi-agent music production pipeline on Amazon Bedrock AgentCore Runtime Instances using shared sessions and GPU infrastructure.
