Amazon Polly launches voice cloning in a limited version (gated release)
Amazon Polly has launched voice cloning in a limited version - from a few minutes of recording it creates a cloned voice in 7 languages. Creation is free, usage is charged on a pay-as-you-go basis; access is still gated for now.
Amazon has launched the voice cloning feature in the Amazon Polly service as generally available, but in a limited version (gated release) - access is already active for selected customers, while others must apply through the AWS account team or request access. The feature can create a cloned voice from a few minutes of audio recordings of a speaker who has given consent, running on the generative engine of Amazon Polly, with the same naturalness and accuracy as the other generative voices. Cloning supports seven languages: English, Spanish, German, French, Italian, Portuguese and Hindi, with each cloned voice speaking only the language in which it was recorded.
According to Amazon, voices are automatically encrypted and protected by access control, so only the customer who created them has access to them. Creating a voice accepts 1-10 recordings (each up to 5 MB) via the console or API (CreateVoice), requires confirmation of TermsAccepted=True as proof of the speaker's consent, and is subject to the AWS Service Terms, the Biometric Notice and Consent Terms, the AWS Responsible AI Policy and the Acceptable Use Policy. Deleting a voice via DeleteVoice removes the data immediately.
The feature is available in all AWS regions that support the Amazon Polly generative engine, including AWS GovCloud (US). Creating a voice is free; usage is charged on a pay-as-you-go basis according to Amazon Polly's pricing. The company cites as use cases consistent brand voice across channels, contact centers, content creation at scale (audiobooks, courses, podcasts), localization with a unified persona, and accessibility/personalization.
Why it matters
The feature allows companies to have one consistent voice across IVR, apps, advertising and content without repeated studio recording, which reduces costs and speeds up script updates. At the same time, it introduces a mandatory confirmation of the speaker's consent and legal terms (Biometric Notice and Consent Terms, Responsible AI Policy), which is relevant for companies dealing with the ethics and compliance of cloned voices. Limited access (gated release) means the feature is not yet automatically available to all customers.
Two audiences, two different impacts
What this means
For individuals
Developers with access to the feature can use the console or API (CreateVoice, GetVoice, SynthesizeSpeech, DeleteVoice) to create a cloned voice in one of 7 languages from 1-10 recordings (max. 5 MB each) and deploy it into their own applications; access is still limited for now and requires confirmation of the speaker's consent.
For a business
Companies can replace repeated studio recording with one consistent cloned voice across IVR systems, mobile apps, advertising, and content (audiobooks, courses, podcasts), which according to Amazon reduces the cost of updating voice prompts and unifies brand voice across all channels. Creating a voice is free; charges apply only once it is used.
Sales and marketingCheck the original
Event sources
clearly official source · 1 publisher, 0 independent. We count feeds from the same owner only once.