Source: aws.amazon.com/blogs/machine-learning/amazon-polly-delivers-a-major-upgrade-to-its-generative-engine
The upgrade aims to make speech output more expressive and accurate, including readback of numbers, currency and addresses. AWS describes it as a drop-in change: voice names and API calls stay the same, so existing applications should keep working. The Generative catalog now totals 57 voices across 11 languages and 25 locales.
The new voices include Kenneth and Nate for US English, Kiara and Arjun for Hindi, and Suvi for Finnish. Hindi and Finnish are generally available on the Generative engine for the first time. AWS also added AWS GovCloud (US-West), bringing the Region count for Generative voices to 11.
Pricing is unchanged, with no new SKU and no cost increase, and cached static prompts can be replayed at no extra charge. Speech marks remain unavailable on Generative voices, while standard and neural voices are unaffected. The Bi-Di streaming API StartSpeechSynthesisStream lets applications send text and receive audio at the same time.
Why it matters if you build agents
Developers already using Polly Generative voices get better output by calling the same API, with no code changes. Teams can add Hindi or Finnish Generative voices to existing voice pipelines, but any flow that depends on speech marks will still need a standard or neural voice. Use DescribeVoices during deployment checks to confirm a chosen voice is available in each Region.
Get stories like this every weekday
One short email with the day's agent tooling news.