Adding custom voices to ElevenLabs means uploading voice samples so the platform can generate speech that sounds like that voice

ElevenLabs lets you create a custom voice by recording or uploading audio samples of a person speaking. Once you upload enough samples, ElevenLabs processes them and adds that voice to your account. You can then use it to generate speech in any text you want. The process takes about 10 to 15 minutes of actual work, though processing can take a few hours.

You need at least one audio file to start, though ElevenLabs recommends uploading multiple samples so the voice sounds more natural and consistent. The samples should be clear, with minimal background noise, and at least a few seconds long each.

Key Takeaways

  • You upload audio files from your computer or record directly in the ElevenLabs interface to create a custom voice.
  • Audio samples should be clear speech with little background noise, and ElevenLabs works best with multiple samples rather than just one.
  • After uploading, ElevenLabs processes the samples for a few hours before the voice appears in your voice list and is ready to use.
  • Once created, your custom voice works the same way as any other voice in ElevenLabs — you select it, paste text, and generate speech.

How to upload audio files for a custom voice

Log into your ElevenLabs account and go to the Voices section. Look for a button labeled "Add a new voice" or a plus icon. Click it, then select "Upload voice samples." A file browser will open.

Choose the audio files from your computer. ElevenLabs accepts MP3, WAV, and M4A files. You can upload one file or multiple files at once. If you are uploading multiple samples, select them all together rather than one at a time — this tells ElevenLabs they belong to the same voice. After selecting your files, click "Upload" or "Open" depending on your browser.

ElevenLabs will ask you to name the voice. Use a clear name you will recognize later, like "Sarah's voice" or "Deep narrator voice." Avoid generic names like "Voice 1" because you may create many voices and will want to find the right one quickly.

Recording audio directly in ElevenLabs

Instead of uploading a file, you can record audio directly in the ElevenLabs interface. In the "Add a new voice" menu, select "Record voice samples" instead of "Upload voice samples." A recording window will open in your browser.

Click the red record button and speak clearly into your microphone. Read a paragraph or two of natural speech — do not rush, and avoid background noise like music or traffic. When you are done, click stop. ElevenLabs will play back what you recorded so you can hear it before saving. If it sounds good, click save. If not, delete it and record again.

You can record multiple samples this way. After each recording, ElevenLabs will ask if you want to record another sample. Click yes, and repeat the process. Three to five samples usually gives ElevenLabs enough to work with, though more is better.

What makes a good voice sample

The quality of your audio files directly affects how well ElevenLabs can create the voice. Speak in a normal, conversational tone — not overly dramatic or robotic. Read naturally, as if you were talking to a friend, not performing. Avoid whispering, shouting, or speaking very quickly.

Record in a quiet room. Background noise like fans, traffic, or other people talking makes it harder for ElevenLabs to isolate the voice you want. If you are recording on your phone, hold it at a consistent distance from your mouth — about 6 to 12 inches. Do not move it around during the recording.

Each sample should be at least 10 to 15 seconds long, though longer is fine. A full paragraph or two is ideal. Avoid samples that are too short, because ElevenLabs needs enough audio to understand the voice's characteristics.

How long processing takes and what happens next

After you upload or record your samples, ElevenLabs processes them in the background. This usually takes 1 to 4 hours, though it can occasionally take longer if the system is busy. You will see a status indicator next to your new voice showing that it is processing.

While processing happens, you can close the browser or work on other things. ElevenLabs will send you a notification when the voice is ready. You can also check the Voices section anytime to see the current status.

Once processing is complete, your custom voice appears in your voice list alongside ElevenLabs' built-in voices. You can now use it like any other voice: paste text into the text box, select your custom voice from the dropdown, and click generate to create speech.

Troubleshooting common problems with custom voices

If your voice sounds robotic or unnatural after processing, the audio samples may have had too much background noise or the speech was too fast or too slow. Try recording new samples in a quieter room and speaking at a normal, conversational pace. Upload those new samples and let ElevenLabs process them again.

If ElevenLabs rejects your files, check that they are in a supported format (MP3, WAV, or M4A). Some audio files from certain apps or devices may not work. Try converting the file using a free tool like Audacity or CloudConvert, then upload the converted version.

If processing takes much longer than 4 hours, refresh the page or log out and back in. Occasionally the status does not update properly. If it still shows processing after 24 hours, contact ElevenLabs support through the help section in your account.

Using your custom voice in projects

Once your voice is ready, select it from the voice dropdown whenever you want to generate speech. The dropdown shows all your custom voices at the top, followed by ElevenLabs' standard voices. Click on your custom voice to select it.

Paste or type the text you want to convert to speech, then click the generate button. ElevenLabs will create an audio file using your custom voice. You can listen to it, download it as an MP3, or use it directly in a video or presentation.

You can create as many custom voices as you want. Each one is stored in your account and available anytime you log in. If you no longer need a voice, you can delete it from the Voices section to keep your list organized.

Frequently Asked Questions

Can I use someone else's voice to create a custom voice?

Technically yes, but you should only do this if you have permission from that person. Using someone's voice without consent may violate their rights. Always get clear permission before uploading audio of someone else.

How many audio samples do I need to upload?

One sample is the minimum, but ElevenLabs recommends at least three to five samples for better results. More samples help the system understand the voice's characteristics and produce more consistent speech. If your first voice sounds off, try uploading additional samples.

What if my audio file is too large to upload?

ElevenLabs has file size limits, usually around 25 MB per file. If your file is larger, compress it using a free tool like Audacity or an online converter. You can also split a long recording into shorter clips and upload them separately.

Can I edit or delete a custom voice after I create it?

You can delete a custom voice from the Voices section anytime. However, you cannot edit the voice itself once it is created. If you want to change how it sounds, you would need to delete it and create a new one with different audio samples.

Do I need a paid ElevenLabs account to create custom voices?

Custom voice creation is available on most ElevenLabs plans, but the free tier may have limits on how many custom voices you can create. Check your account plan to see what is included, or upgrade if you need more custom voices.