Amazon Polly voice cloning is generally available as a gated release in seven languages
Amazon Polly voice cloning is now generally available as a gated release. It builds a custom voice on the Polly generative engine from a few minutes of one consenting speaker's audio, in seven languages. Building a voice is free, and usage is billed pay-as-you-go.
Amazon Polly voice cloning is now generally available as a gated release, AWS announced on its Machine Learning Blog. Given a few minutes of audio from a speaker who has agreed to be cloned, Polly builds a custom voice that runs on the Amazon Polly generative engine. AWS says these cloned voices deliver similar naturalness and accuracy to the other generative voices in Polly.
Languages, access and pricing
- Languages: English, Spanish, German, French, Italian, Portuguese and Hindi. Each cloned voice speaks only the language it was recorded in.
- Regions: All AWS Regions that support the Polly generative engine, including AWS GovCloud (US).
- Access: Already enabled for select Polly customers. Others can contact their AWS account team or apply for access.
- Pricing: Building a cloned voice is free. Cloned voice usage is charged pay-as-you-go. Specific rates are listed on the Amazon Polly Pricing page.
How the lifecycle works
The workflow has four steps, all exposed through the API and console.
- Record. Record one speaker. Recordings can be sent directly in the CreateVoice request, uploaded in the console, or stored in an Amazon S3 bucket in the AWS Region where the voice will live.
- Create. Call CreateVoice with 1-10 recordings, the recording language, and TermsAccepted=True, which confirms you hold the speaker's consent. Polly returns a voice ID (vc-...) immediately and builds the voice in the background.
- Wait. Creation is asynchronous and typically takes 5-10 minutes. Polly also checks the TTS voice for speaker similarity during this time. Call GetVoice until the status moves from pending to active or failed.
- Use. Pass the voice ID to SynthesizeSpeech for real-time audio, or to StartSpeechSynthesisTask for long-form audio written to S3, with Engine="generative". DeleteVoice removes the voice.
Consent, security and data handling
Consent is confirmed each time a voice is created, in the console or through TermsAccepted=True in the API. That confirmation covers having obtained all rights and consents needed to create and use the voice. Use is subject to the AWS Service Terms, the Biometric Notice and Consent Terms, the AWS Responsible AI Policy and the AWS Acceptable Use Policy.
Audio generated with a cloned voice carries a watermark. A cloned voice is scoped to the AWS account that created it, with access controlled through IAM. Voices can be encrypted with a customer-managed AWS KMS key and tagged.
According to AWS, Polly does not keep a copy of the voice recordings once the voice is created and does not use them to train Polly models. It retains only the encrypted voice data needed to synthesize speech, and DeleteVoice removes that data immediately. AWS recommends keeping a record of each speaker's consent.
Intended uses
AWS lists these use cases:
- A consistent brand voice across IVR systems, cascaded voice agents, mobile apps, in-store announcements, advertising and connected devices.
- Contact center prompts that can be updated without a new studio session.
- Audiobooks, courses, podcasts, game characters and videos.
- Localization with a consistent persona. For multilingual experiences, AWS says to create and review approved voices per supported language.
- Accessibility and personalization, including reading experiences and assistive tools.
The announcement includes audio demos from AWS staff, comparing a real recording with the cloned voice reading new text. One is a German sample, and another is a support-line prompt.
What this means
Voice cloning is exposed as a set of API calls (CreateVoice, GetVoice, DeleteVoice) alongside the existing synthesis calls, so teams can build voice creation and deletion into their own AWS applications. Cloned voices are synthesized the same way as other generative voices, by setting Engine="generative".
Because each cloned voice is tied to one recording language, a brand that wants the same persona in several of the seven supported languages needs a separate voice per language, each with its own recordings and consent confirmation.
The controls are built around the speaker's consent and the customer's ownership of the voice. These include per-voice consent confirmation, account-scoped access, optional customer-managed KMS keys, watermarked output, and immediate deletion of voice data. Access remains gated, so most teams will need to go through their AWS account team or apply before they can create voices.
Related Articles
Kimi Code CLI v2.7.0 adds native Amazon Bedrock provider with standard AWS credentials
Kimi Code CLI v2.7.0, Moonshot AI's open source coding agent, adds a native bedrock provider. It resolves the standard AWS credential chain, including named profiles and SSO, so users no longer need a separate Amazon Bedrock API key. AWS demonstrates the setup with Kimi K3.1 Pro Highspeed.
Nous Research brings open-source Hermes Agent to Microsoft Teams, adds Collective Wisdom skill sharing
Nous Research says companies will soon be able to run its open-source Hermes Agent directly in Microsoft Teams, with agents sandboxed in Azure Container Apps. A new feature, Collective Wisdom, lets employees share reviewed skills with colleagues, and the shared knowledge stays inside the company.
BioMap launches AI drug safety system with xTrimoTox, reports 0.9637 AUROC on liver injury
BioMap launched an AI-assisted nonclinical drug safety assessment system on November 19. It combines virtual-cell models, toxicity prediction and evaluation services, and the company reports AUROC of 0.9637 for drug-induced liver injury on a selected benchmark task.
OpenAI tests ChatGPT dots with a small group of Plus users, rolls out to all Business Standard
OpenAI has started testing ChatGPT dots with a small number of ChatGPT Plus subscribers and is rolling it out to all ChatGPT Business Standard users. Dots launched in September for Pro, Business Premium and Enterprise plans. Usage allowances for the new tiers have not yet been clarified.
Comments
Loading...