Privacy
Your users' voices are not our data
Voice is personal. We process it to produce structured data for you, then we delete it. Here's exactly what happens today and what we plan.
Draft — not a legal policy
This page describes the prototype's behaviour in plain language. A formal privacy policy will replace it before launch.
| Topic | Today (prototype) | Planned for production |
|---|---|---|
| Voice recordings | Processed and deleted after each request. Not stored, not used for training. | Same default. Any retention would be opt-in, time-limited and documented. |
| Transcripts & fields | Returned to the caller only. Not kept by the Voice API. | Optional request history in the dashboard, off by default, with a retention setting. |
| Model training | Customer audio is not used to train models. | Unchanged unless a customer explicitly opts in. |
| This website | The live demo sends your recording to the Voice API you're connected to; the demo-request form is validated but not stored. | Contact data stored in a CRM to answer your request; analytics only with consent. |
| Sub-processors | None — models run locally. | A published list of any hosting providers used for the managed service. |
You stay in control
When you integrate Nabra, you remain responsible for informing your users that their voice is processed, and for what you do with the structured result. The API itself makes no decisions and contacts no one.
Questions
Email privacy@example.com (placeholder address).