Amazon Polly Voice Cloning Now Generally Available: 7 Languages, Free Voice Creation, Access by Request
AWS announced that Amazon Polly voice cloning is generally available as a gated release. It supports 7 languages, voice creation is free and usage is pay-as-you-go. Access is enabled per account; request it through your AWS account team or the request page.
News date: About 6 min read
Contents
AWS announced that voice cloning for Amazon Polly, its text-to-speech service, is generally available, meaning it is a finished product rather than a preview. It is a gated release, though: it has been enabled for some Polly customers, and others must contact their AWS account team or submit a request. AWS says businesses and developers can use one brand voice for customer service prompts, audiobooks and app narration, without re-recording every time a script changes.
What is Amazon Polly voice cloning, and how does it differ from standard voices?
According to AWS, a customer provides a few minutes of recordings from a speaker who has consented to having their voice cloned, and Polly creates a dedicated cloned voice. The voice runs on Polly's generative engine, the AI technology Polly uses to produce more natural speech. AWS says its naturalness and accuracy are comparable to other generative voices. Cloned voices are automatically encrypted and protected by access controls, and AWS says only the account that created a voice can use it.
Uses listed by AWS include: a consistent brand voice across IVR (interactive voice response phone systems), voice agents, mobile apps and connected devices; contact center prompts; audiobooks, courses, podcasts and game characters; and accessible reading and personalized interactions.
What stage is it at, and what languages, Regions and limits apply?
AWS describes this as a generally available, gated release. Supported Regions are the same as for Polly's generative engine, including AWS GovCloud (US). The Regions AWS lists are US East (N. Virginia), US West (Oregon), Frankfurt, Zurich, London, Sydney, Seoul, Singapore, Tokyo, Canada (Central) and GovCloud (US-West).
| Item | What AWS states |
|---|---|
| Release stage | Generally available, gated release, enabled per account |
| Languages | English, Spanish, German, French, Italian, Portuguese, Hindi |
| Recordings | 1–10 clips per voice, up to 15 minutes in total; 2–10 minutes recommended |
| Creation time | Typically 5–10 minutes, asynchronous (completes in the background) |
| Quota | Initial 10 cloned voices per account per Region; CreateVoice and DeleteVoice share a limit of 1 request per second |
| Cost | Voice creation is free; usage is pay-as-you-go, see the Polly pricing page |
| Controls | Tied to the creating account; managed through IAM permissions; can be encrypted with your own KMS key |
How do you get started?
AWS explains that if your account is enabled, "Voice cloning" appears in the left panel of the Polly console. The console steps AWS lists are:
- In the console, expand the left panel and choose Voice cloning.
- Upload up to 10 recordings, or select them from Amazon S3; each console upload is limited to 5 MB.
- Confirm the terms and choose Create.
- When the voice is ready, preview it on the details page.
To use it programmatically, AWS says you need the latest version of boto3 (the AWS SDK for Python) and the relevant IAM permissions. IAM is AWS's account permissions management service. The flow is to call CreateVoice with TermsAccepted=True, check status with GetVoice, then call SynthesizeSpeech or StartSpeechSynthesisTask with Engine="generative" to produce speech. See the original AWS post for full code examples.
What does it mean for real workflows?
AWS recommends treating voice creation as a preparation step rather than part of a user request, because creation takes a few minutes. For fixed prompts such as phone greetings, AWS suggests generating the audio ahead of time, having the speaker and brand team approve it, and then serving the stored file, because the same text can vary slightly between requests. AWS also notes that voice style is taken from the recordings and cannot blend multiple emotions, and recommends testing with your own product names, numbers and addresses. Phoneme tags, markup in SSML (Speech Synthesis Markup Language) that spells out how a word should be pronounced, work with cloned voices in English, French, Italian, German and Spanish.
On data handling, AWS says Polly does not retain a copy of the recordings after the voice is created and does not use them to train Polly models. Polly keeps only the encrypted voice data needed for synthesis, which can be encrypted with the customer's own KMS (AWS Key Management Service) key, and calling DeleteVoice removes it immediately.
Frequently asked questions
Can all AWS customers use it now?
No. AWS says access is enabled per account. Some Polly customers are already enabled; others need to contact their AWS account team or request access.
Can one cloned voice speak multiple languages?
According to AWS, each cloned voice speaks only the language used in its recording. For a multilingual experience, you need to create and review a separate voice for each supported language.
How much does it cost to create a voice?
AWS says creating a cloned voice is free and usage is charged on a pay-as-you-go basis; check the Amazon Polly pricing page for specific rates.
How many cloned voices can I create?
AWS says the initial quota is 10 per account per Region; to raise it, contact your account team or AWS Support.
Can generated audio be identified as synthetic?
AWS says Polly adds a watermark to the text-to-speech audio produced by cloned voices.
Browse the latest news in this topic
Found this guide helpful? Share it with someone who needs it
Claude Code 2.1.329: Code Mode Works With Telemetry, /tasks Lists Scheduled Tasks, macOS Computer Use Gets Stricter Safety Rules
Anthropic's changelog lists Claude Code 2.1.329, covering Code Mode behavior changes, new diagnostics in /tasks and /mcp, /code-review changes, and new safety rules for macOS computer use.
Kimi Code CLI v2.7.0 Adds a Native Amazon Bedrock Provider: Use Your AWS Credentials Instead of a Separate API Key
AWS Rounds Up 13 SageMaker Inference Features: Managed NVIDIA Dynamo on HyperPod, MIG GPU Partitioning, Cross-Region Routing
BMW Group Shares How It Governs AI Agents: 12,500+ Developers, Three Autonomy Levels and a Four-Stage CI/CD Gate
Sources
- Amazon Polly launches voice cloning: A unique voice for your brand | Artificial IntelligenceChecked:
Compiled from the references and verification dates listed above; fares, schedules, product features, and terms can change, so please confirm the latest details on the relevant official website before travelling, booking, or purchasing.