Amazon Polly delivers a major upgrade to its Generative engine

Amazon Polly’s Generative engine has been upgraded to deliver more expressive and accurate speech with low latency, adding 14 new voices across 9 locales and making Hindi and Finnish generally available for the first time. The catalog now totals 57 generative voices spanning 11 languages and 25 locales, while existing voice names and API calls remain unchanged for a drop‑in upgrade. Engineers can immediately leverage the richer voice selection and improved prosody handling for LLM‑based agentic applications, contact centers, and media workloads.

⚡ Key Takeaways

• 14 new voices added across 9 locales, including first‑time GA for Hindi and Finnish.
• Drop‑in upgrade: voice names and API calls are unchanged, enabling seamless migration.
• Improved naturalness and expressive prosody based on textual context without manual configuration.
• Total catalog now 57 generative voices covering 11 languages and 25 locales.
• Supports accurate rendering of numbers, currency, dates, and addresses (e.g., “4471 9902”, “$1,284.50”).

The expanded, higher‑quality voice roster lets production systems replace static TTS with dynamic, context‑aware speech, reducing latency and engineering effort while enhancing user experience in voice‑driven applications.

✅ Practical Steps

1. Update your application to point to the latest Amazon Polly Generative engine endpoint (no code changes required).
2. Test the new voices using the provided sample audio URLs to evaluate naturalness and prosody for your use cases.
3. Replace existing voice identifiers with the new voice names (e.g., “Kenneth”, “Kiara”, “Suvi”) as needed to leverage the expanded catalog.

Want the full story? Read the original article.