How do I clone my voice?

How does cloning work?

• You read a consent sentence Audexum generates for you, in your language. It is available in English, German, French, Spanish, Italian, Portuguese and Russian; other languages read the English sentence.
• Audexum checks with speech-to-text that you actually said it, then builds a voice print from the recording.
• The recording must be at least 6 seconds and at most 60 seconds. Uploaded files can be up to 12 MB.

Your cloned voice then appears in Studio next to the other voices.

Whose voice may I clone?

Your own voice, or the voice of a person who gave you documented permission and who reads the consent sentence themselves. Passing the check is not permission. The check only confirms someone read the sentence.

You may not clone public figures without permission, voices from podcasts, films or phone calls you did not make for this purpose, deceased people without documented rights, or any child. Never use a cloned voice to make a real person appear to say something they did not. See the Acceptable Use Policy, section 2.

How is a cloned voice used and stored?

Cloned voices are used in Studio and the API like any other voice. Generation normally costs the V3 rate of 2 credits per character. AI assistants connected through MCP cannot use cloned voices.

Audexum keeps the voice print and consent record until you delete the voice or close your account. The stored reference recording is deleted 12 months after enrollment. On the free plan, deleting your one clone does not free up the slot.