> ## Documentation Index
> Fetch the complete documentation index at: https://docs.deepblocker.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Voice cloning and consent

> The rule for cloning a named person's voice, what the person records, what we store, how they withdraw, and what the platform refuses.

A simulation uses a cloned voice of a named person only after that person has given recorded consent. We never clone a voice from a public page or a recording the person did not give for this purpose. The six stock voices need no consent.

For the steps in the console, see [Executive voice consent](https://docs.deepblocker.ai/vishing/executive-voice-consent.md).

## How consent is recorded

| Step | What happens |
| - | - |
| 1 | Your team enters the executive's name, role and email. They get a one-use link by email |
| 2 | The link is valid for 7 days |
| 3 | The executive reads five short points: what a clone is, what it is used for, what is stored, how to withdraw, and who to contact |
| 4 | They tick that they have read the points and record one sentence in their own voice |
| 5 | They record about two minutes of speech (at least 60 seconds), or upload up to three audio files instead |
| 6 | They submit. The clone is made and the consent record is stored with it |

The sentence every executive reads aloud:

> My name is (your name). I agree that (your company) may make a copy of my voice and use it in security awareness tests of its own staff. I know I can withdraw this at any time.

## What is stored

| Item | Why |
| - | - |
| The two minutes of speech | To build the clone |
| The recording of the spoken consent sentence | To show the voice was made with the person's agreement |
| Their name, role and email address | To identify whose voice it is and reach them |

The licence treats voice samples as biometric data. See [Legal documents](https://docs.deepblocker.ai/trust/legal.md).

## Withdrawal

The executive opens the same link and chooses **Withdraw my consent**. Then:

| What | Happens |
| - | - |
| The clone | Deleted at the voice provider |
| The speech sample | Deleted from storage |
| Campaigns that use the voice | Blocked. The campaign page says who withdrew, on what date, and asks for another voice |
| The consent recording | Kept, with the time of withdrawal, as the record of what was agreed |

Withdrawal cannot be undone. A new copy needs a new consent.

## Request statuses

| Status | Meaning |
| - | - |
| **sent** | The link has gone out and has not been used |
| **recording** | The executive has submitted and the clone is being made |
| **recorded** | Consent is on file and the voice is ready |
| **revoked** | The executive withdrew. The clone is gone |
| **expired** | The link was not used within 7 days. Send a new one |

## What the platform refuses

| Rule | Detail |
| - | - |
| No consent, no launch | A campaign whose voice has no consent record cannot be launched |
| Withdrawn voice | A campaign pinned to a withdrawn voice is blocked until another voice is chosen |
| Executive voices per workspace | 3 by default. Open invitations count towards the limit. Ask us to raise it |
| One use per link | A link that has been used to record cannot record again |

## Deletion

A cloned voice and its sample are deleted 30 days after the voice is created. See [Retention and deletion](https://docs.deepblocker.ai/trust/retention.md). The licence states the same rule as: cloned only with the written consent of its owner, for the named campaign, and deleted within 30 days of the report.

> [!NOTE]
>
> An executive with a question about their data can write to [privacy@deepblocker.ai](mailto:privacy@deepblocker.ai). That address is on the consent page they see.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.