Security & Compliance
Your data never trains our models
We never use your audio, job or stream outputs, or any other customer data to train, improve or develop our models. This holds on every plan, with no opt-out needed.
We are a European company
pyannote is a French company governed by French and EU law. We develop our models in France and host our core services in the European Economic Area.
You decide where processing runs
Use our hosted API, or run our models in your own cloud or on-premises. For the strictest requirements, your audio never leaves your infrastructure.
Data Lifecycle & Usage
We don't train on customer audio. Many speech AI providers train on customer data by default and ask you to opt out, or reserve that protection for higher-priced plans. We don't. The commitment is part of our standard contract.
Data sent to pyannoteAI is processed for a single purpose: running the speaker intelligence job or stream you requested. We do not analyse, aggregate or share it for any other use.
If you report an issue with accuracy or unexpected output, you can choose to send us example data. We use that data for validation and testing only with your explicit consent, and only for the scope of your report.
We keep data only as long as processing requires it. What happens to your file depends on how you submit it.
Audio on processing servers is deleted as soon as the job completes, whether you submit a URL or an uploaded file.
Files uploaded through our Media API stay in temporary storage for 48 hours, so you can reuse them across jobs. They are then permanently deleted.
Streamed audio and stream outputs are never stored.
Job outputs, including diarization, identification, transcription segments and generated voiceprints, are deleted 24 hours after the job completes.
Voiceprints submitted for identification have their content redacted within 96 hours.
Encryption of data at rest and in transit
Host hardening and infrastructure managed as code
Restricted access to production deployments
Regular penetration testing
Confidentiality obligations for every employee and contractor who processes personal data
Compliance
GDPR
pyannoteAI processes personal data in accordance with the GDPR and the French Data Protection Act. A Data Processing Agreement is included in our terms of use for every plan, and we support data subject requests for access, rectification, erasure, and portability.
HIPAA
pyannoteAI supports teams processing protected health information. Business Associate Agreements are available on Enterprise plans.
SOC 2 Type II
Our audit is in progress. The report will be published on the Trust Center once issued.
Does pyannoteAI use my audio or outputs to train its models?
Yes, absolutely! You can sign up for a free trial and enjoy 150 hours of identification and 10 voiceprints for one month. You will have access to our latest premium models to ensure accuracy and precision for your speech processing needs.
How long do you keep my audio files?
Yes, absolutely! You can sign up for a free trial and enjoy 150 hours of identification and 10 voiceprints for one month. You will have access to our latest premium models to ensure accuracy and precision for your speech processing needs.
Where is my data processed?
Yes, absolutely! You can sign up for a free trial and enjoy 150 hours of identification and 10 voiceprints for one month. You will have access to our latest premium models to ensure accuracy and precision for your speech processing needs.
How do you handle GDPR data subject requests?
Yes, absolutely! You can sign up for a free trial and enjoy 150 hours of identification and 10 voiceprints for one month. You will have access to our latest premium models to ensure accuracy and precision for your speech processing needs.
Where can I review your compliance and security documentation?
Yes, absolutely! You can sign up for a free trial and enjoy 150 hours of identification and 10 voiceprints for one month. You will have access to our latest premium models to ensure accuracy and precision for your speech processing needs.
Where can I access your terms and conditions?
Yes, absolutely! You can sign up for a free trial and enjoy 150 hours of identification and 10 voiceprints for one month. You will have access to our latest premium models to ensure accuracy and precision for your speech processing needs.