Voice Cloning

Voice Cloning lets you create a custom voice for your AI receptionist by uploading an audio recording or recording one directly in your browser. The cloned voice can then be selected in the Voice Library and used on calls.

Where to Find It

  1. In your dashboard, navigate to Voice Library.
  2. Select the Clone Voice tab at the top of the page.

Creating a Voice Clone

The Create Custom Voice form has four fields and an audio source selector.

1. Voice Name

Enter a name for your custom voice (required). This is how the voice will appear in your library.

2. Description

Provide a short description of the voice (required). For example, "Warm and professional female voice for customer service."

3. Voice Provider

Choose which provider processes the clone. Select one of:

  • Elevenlabs
  • Cartesia

Click the provider name to toggle between them.

4. Audio Source

Provide the voice sample using one of two methods:

  • Select File -- Click to upload a .wav or .mp3 file from your computer (maximum 10 MB).
  • Start Recording -- Click to record directly using your microphone. The button turns red and pulses while recording. Click Stop Recording when finished.

After uploading or recording, you will see a confirmation:

  • Uploaded files display the file name.
  • Recordings display an audio player so you can review the clip. Click the X button to discard and re-record.

5. Submit

Click Create Voice Clone to upload your audio and begin the cloning process. A success message confirms when the voice has been created and attached to your receptionist.

Your Cloned Voices

After creating one or more clones, a Your Cloned Voices section appears below the creation form. Each cloned voice is shown as a card, just like voices in the main library.

From these cards you can:

  • Preview -- Play a sample of the cloned voice
  • Select -- Apply the voice to your receptionist
  • Delete -- Remove the cloned voice

Requirements and Limits

  • Supported formats: WAV, MP3
  • Maximum file size: 10 MB
  • For best results, use a clear recording with minimal background noise
  • Speak naturally for at least 10--15 seconds to give the provider enough data

Next Steps