Confirm the authorized use
Use your own voice or an explicit authorization from the speaker. Agree on the intended script and context. Do not use cloning for fraud, harmful impersonation, or deceptive identity claims.
Audio / Audio & Voice Cloning
A voice carries identity.
Treat it that way.
Prepare a permission-based voice workflow with an authorized reference, a clear script, and a careful listening review.
A voice is an identifiable performance, not simply a sound preset. Start with explicit permission and a clean reference. Listen for articulation, pace, tone, and the moments where the delivery no longer feels consistent with the authorized source.
Your inputs, directions, and versions stay together in Studio.
Voice Cloning / Voice studyResult pending
A voice, with permission.
Missing audio sampleA verified recording has not been added yet.
Choose your starting point
Keep the source and the purpose clear.
Choose the workflow based on whether you need a voice identity, new sound, or a continuation.
A considered workflow
Authorize. Prepare. Listen.
A technically clean sample does not replace permission.
Choose a clear reference
Use a recording with intelligible speech and minimal competing sound. Follow the connected workflow’s actual input requirements when they are available.
Review identity and meaning
Listen for pronunciation, pacing, unnatural artifacts, and unintended changes to meaning. Make sure the final context does not misrepresent what the speaker said or approved.
Make the brief do the work
Direct the performance, not deception.
A production brief should record the authorized intention as well as the desired delivery.
Example direction
Use the speaker-authorized voice reference for the approved tutorial script. Keep the delivery clear, measured, and conversational. Preserve the supplied wording, and review the result with the speaker before publication.
This is an example creative brief, not a record of the sample media’s generation settings.
01
A voice reference is not a blanket permission
Access to a public recording does not establish permission to clone the speaker. The intended use matters, including the script, the audience, and whether the result could be mistaken for an authentic recording.
Use an authorization process appropriate to the project. The website’s terms and guidelines still apply; this workflow guide does not grant permission to use another person’s voice.
02
Listen to more than similarity
A result can resemble a voice while mispronouncing a name, changing emphasis, or introducing artifacts. Similarity alone does not make the recording suitable for use.
Review the wording against the approved script and listen in the final mix. Obtain speaker review when that is part of the permission or production process.
03
The demonstration needs both sides
A useful example pairs an authorized reference with the generated result and a transcript. It should say what was supplied and what the tool produced.
Those recordings are currently missing. The page will not substitute an unrelated sample or a simulated waveform for a verified cloning demonstration.
The useful details
A few good
questions.
Can I use a public figure’s voice recording?
A publicly accessible recording is not permission to clone the speaker. Use an authorized source and a legitimate, non-deceptive purpose.
Does the redesign add a cloning backend?
Open this workflow in Studio to prepare its inputs, direction, and project assets. Review the supported options and the service quote before generating.
What should the source recording contain?
Clear, authorized speech without distracting competing sound. The exact format and sample requirements must come from the connected workflow.
Why should I provide a transcript with the sample?
It lets reviewers compare the intended wording with the actual output and helps people who cannot listen to the recording.
Connected workflows
Keep the idea moving.
Your next step
Take the idea further.
Explore the possibilities here, then continue in the existing Studio.