As part of Microsoft’s commitment to responsible AI, custom neural voice is designed with the intention of protecting the rights of individuals and society, fostering transparent human–computer interaction, and counteracting the proliferation of harmful deepfakes and misleading content. For this reason, custom neural voice is a Limited Access feature available by registration only, and only for certain use cases.
Approved Limited Access registrants for custom neural voice will also get access to deploy and use a Custom Neural Voice Lite voice model. Custom Neural Voice Lite is a project type in public preview that allows you to record 20-50 voice samples on Speech Studio and create a lightweight custom voice model for demonstration and evaluation purposes. Both the recording script and the testing script are pre-defined by Microsoft. A synthetic voice model you create using Custom Neural Voice Lite may be deployed and used more broadly only if you apply for and receive full access to custom neural voice (subject to applicable terms).
Personal voice (preview), including the try-with-your-own-data demo experience in Speech Studio, is a Limited Access feature and requires registration and approval. Customers must register to access the demo and/or the API for business use.
As a Limited Access feature, access to custom neural voice requires registration. Only customers managed by Microsoft, meaning those who are working directly with Microsoft account teams, are eligible for access. Customers who wish to use this feature are required to register by submitting the Limited Access registration form. Your use of custom neural voice is limited to the use case(s) that you select and Microsoft approves at the time of registration. Microsoft may require customers to re-verify information submitted for registration.
Custom neural voice is made available to customers under the terms governing their subscription to Microsoft Azure Services (including the Service Specific Terms). Please review these terms carefully as they contain important conditions and obligations governing your use of custom neural voice. For example, these terms include, but are not limited to, the following obligations:
- Voice talent and approved use cases. Customers must warrant that they have obtained explicit written permission from voice talent prior to creating a voice model, and must share the Disclosure for voice and avatar talent with the voice talent in advance. Customer may use each custom voice model only for use cases approved during registration.
- Implementation requirements. As outlined in our Code of conduct, in addition to other requirements, customers must not use any custom voice model for prohibited uses and must also agree that when deploying each custom voice model, they will disclose the synthetic nature of the service to users and support a feedback channel that allows users of the service to report issues and share details with Microsoft.
- Microsoft's additional processing and use of voice talent data. Before a customer can train a custom voice model, Microsoft will require the customer to upload a recorded audio file to the Speech Studio with a pre-defined statement from the voice talent acknowledging that the customer will use the talent's voice to create a synthetic voice. As a technical safeguard intended to help prevent misuse of this service, Microsoft reserves the right to use Microsoft's speaker recognition biometric identification technology on this recorded acknowledgement statement and verify it against the training audio data to confirm that the voices are from the same speaker. Microsoft will also continue to retain this recorded acknowledgement statement file and the custom voice model to protect the security and integrity of our services. Customer is responsible for ensuring all necessary permissions are obtained from voice talent for these purposes.
Help and support
FAQ about Limited Access features can be found here.
If you need help with custom neural voice, find support here.
Report abuse of custom neural voice here.