How to clone your voice in ElevenLabs
Prepare a clean, authorized recording, choose the right cloning flow, and check the result before using it in narration.
Updated 5 min read
To clone your voice in ElevenLabs, prepare a clean recording and choose either Instant Voice Cloning or Professional Voice Cloning. ElevenLabs documents separate creation flows (opens in a new tab): Instant includes a rights-and-consent confirmation, while Professional requires verification of your own voice. Keep the sample conditions and review findings in fenbs, so you can continue the work in another AI without explaining which recording you accepted.
Which cloning option should you choose?
Choose the flow before recording a large set of samples. A short test and a dedicated training recording are different jobs. Check the current ElevenLabs plan details (opens in a new tab) for access rather than assuming every account includes both options.
| Flow | Preparation | What to account for |
|---|---|---|
| Instant Voice Cloning | A short, clean, consistent sample | Confirm that you have the right and consent to clone the voice. |
| Professional Voice Cloning | A considerably larger set of recordings of your own voice | Complete voice verification and allow for processing. |
The Professional Voice Cloning guide (opens in a new tab) is explicit: this flow is for your own voice. Someone else agreeing to be cloned does not make their voice eligible for your Professional clone. Follow the guide’s current recording recommendations and exclude sung training material.
Where do you create the voice?
- For Instant Voice Cloning, open Voices in ElevenLabs and select the plus icon.
- Choose Instant Voice Clone, then follow the instructions to upload or record your authorized audio.
- Name the voice and confirm the rights-and-consent statement. Select Save voice.
- For Professional Voice Cloning, use the separate Voices > Create Voice > Professional Voice Clone flow.
- Upload your own samples and complete the voice verification instructions. Allow the Professional clone to finish processing before assessing its output.
Use a name you can understand later, such as “My tutorial narration, quiet-room sample.” In your own notes, distinguish the recording date from the date you reviewed the clone. Creating a voice and approving it for a particular piece of work are separate milestones.
What should the source recording sound like?
The Instant guide recommends a short stretch of clean, consistent audio. Prioritize a steady performance over collecting many unrelated clips. Record in the speaking style you actually want to assess, rather than combining a whispered introduction, an excited announcement, and a distant conference recording.
- Choose a quiet recording location and listen for background sound before beginning.
- Keep your microphone position consistent across the samples you plan to compare.
- Use a comfortable delivery that resembles the narration you intend to make.
- Listen through the entire sample, including its beginning, pauses, and ending.
- Keep an untouched original and a written description of any preparation you applied.
Do not treat fan noise, room echo, or accidental microphone movement as part of a desired voice style. If the source needs attention, separate that job from cloning; the Voice Isolator preparation and review guide explains how to compare a cleanup pass without losing its source.
How do you check whether the clone is usable?
Write a short review script before testing. Include an ordinary sentence, a sentence with a name you need, and a longer sentence with a natural pause. Use the same script when comparing candidates, so a more appealing phrase does not distract you from the voice itself.
- Listen to the original sample first and describe the delivery in plain words.
- Listen to the candidate at a comfortable level, using the same playback setup.
- Check whether the pacing, tone, and word endings suit the intended narration.
- Mark the exact sentence or timestamp where something needs review.
- Record whether you accept the candidate, want another test, or need a better source recording.
For example, “clear on the opening sentence; the product name still needs review” gives the next session a useful task. “Sounds good” does not. If the destination is a slide deck, the PowerPoint narration guide covers that script-planning work separately.
What if verification or the sound needs more work?
First identify what failed. An incomplete Professional verification, a clone still processing, and a finished voice that does not suit your script need different next steps. Follow the Professional guide for verification and processing; do not keep changing the creative brief while trying to finish an account workflow.
When reviewing an unsatisfactory sound, return to the source and your listening notes. Change one preparation choice at a time and keep the same review script. That produces a comparison you can explain, even when you decide the recording needs to be made again.
Keep the recording review separate from later work such as turning audio into checked notes or preparing meeting follow-ups. Those workflows answer questions about spoken content; they do not establish that a cloned voice is ready for publication.
What should you keep after approving a test?
Keep the permitted source location, recording conditions, chosen voice name, review script, and your acceptance notes. Write down which model and settings you actually used when that information is available. Treat this as a recipe for another review, rather than a promise of identical speech.
ElevenLabs does not export voice clones as standalone files (opens in a new tab). Its supported alternative is generating speech through the ElevenLabs API while the voice stays in the account. Preserve your source samples separately: recreating a clone can sound different even with the same audio.
The useful thing to carry between AI apps is the checked work record: what you tried, what you accepted, and what remains uncertain. The AI at work guides cover keeping those outcomes useful after a single session ends.
Never lose the work you did in an AI session
After a useful AI session, save it in fenbs: the key observations, findings, research and context. Then carry on from exactly where you left off, on any device and in any other AI. The next AI you connect reads it at the start of the session, so you never have to explain the same problem again.
Create a free accountFree for up to 5 people and 50 open tasks · No credit card · Works with Claude, ChatGPT, Cursor and more
How to keep your voice review with fenbs
- Create a free account.
- Write what the voice is for and add a link to the source recording only if you are allowed to share it.
- Save the recording conditions, the review sentences, and which candidate you accepted. Add the next thing to check.
- Connect ElevenLabs to fenbs through ElevenAgents if you use its assistant tools. Save your findings yourself when working in Studio.
- Before you stop, write where you left off. Open that record on your next device or ask another connected AI to read it before helping.
ElevenAgents’ external-tool connection (opens in a new tab) uses MCP, a standard way for an assistant to use an outside tool. It requires workspace opt-in and is unavailable for Zero Retention Mode or HIPAA-enabled workspaces. This does not establish a Studio connection or automatic transfer of recordings. fenbs keeps the work notes; it does not create, verify, or host your cloned voice.