Skip to content

Amazon Polly's Generative voice engine adds Hindi and Finnish, plus GovCloud

· by Pondero Newsdesk

The short version

AWS swapped in a new model behind Amazon Polly's Generative text-to-speech engine, adding 14 voices across nine locales and bringing Hindi and Finnish onto the engine for the first time.

Amazon Polly's Generative voice engine adds Hindi and Finnish, plus GovCloud

AWS swapped in a new model behind Amazon Polly's Generative text-to-speech engine on October 20, bringing Hindi and Finnish onto that engine for the first time and pushing the whole lineup into AWS GovCloud, per AWS's What's New post.

What AWS shipped

The upgrade adds 14 new voices spread across nine locales: Kenneth and Nate for US English, Russell and Curtis for Australian English, Kiara and Arjun for Hindi and Indian English, Tina for German, Carlos for US Spanish, Leo and Carolina for Brazilian Portuguese, Hyunduk for Korean, and Suvi for Finnish, per the post. Hindi and Finnish are new to the Generative engine specifically; Polly already covered other languages through its older, non-generative engines. AWS says the underlying model change also makes tricky content, such as numbers, email addresses, and street addresses, sound more reliable without manual tuning. SSML tags still work for output control, so existing integrations do not need a rewrite. The new voices land in eleven AWS Regions, including US East, US West, three European regions, four Asia Pacific regions, Canada Central, and AWS GovCloud (US-West), per the same source.

Polly's Generative engine is AWS's low-latency, higher-naturalness text-to-speech tier, positioned for contact centers, chatbots, and real-time voice agents built on large language models. The GovCloud rollout is the detail worth flagging separately: it brings the Generative voices to AWS's regulated and public-sector cloud for the first time, which matters for any agency or contractor that previously had to fall back to Polly's older, less natural-sounding engines to stay in a compliant Region.

A coverage gap that specifically touches agentic voice apps

For anyone building a voice agent that needs to serve Hindi or Finnish speakers, the practical change is that they no longer need to patch together a lower-latency engine for most languages and bolt on Polly's older, choppier engine wherever Hindi or Finnish users show up. AWS's low-latency streaming API is aimed squarely at real-time use, so a single consistent voice quality across languages removes an integration compromise that multilingual contact-center and chatbot builders previously had to design around. The GovCloud availability does similar work for regulated builders: a government agency's voice assistant or IVR system can now use the same Generative voices as a commercial deployment, rather than a degraded substitute.

Hindi in particular is a large underserved market for generative voice. India has hundreds of millions of Hindi speakers, and most major text-to-speech vendors have historically prioritized English, Spanish, and a handful of European languages before expanding further. AWS reaching Hindi before some competitors have is a specific, checkable claim about coverage, not a quality claim; AWS does not publish a comparative naturalness benchmark against ElevenLabs or OpenAI in the announcement.

What to watch next

Watch whether ElevenLabs or OpenAI respond with their own Hindi or Finnish releases on comparable low-latency tiers, since both languages have been comparatively thin spots in the major vendors' voice-model coverage. Also worth tracking is whether AWS extends GovCloud availability to future Generative voice additions as a standing policy rather than a one-time catch-up.

Sources