Skip to main content
Clone Voices (also known as Custom Voices) allow you to train a personal voice model that you can use across all your projects. Once created, you can generate speech with this voice repeatedly using the Advanced Model.
Clone Voices result in a unique Voice ID saved to your account.

​
How to Create a Voice

1

Navigate to Clone Voices

Log in to your account and go to the Clone Voices page.
2

Provide Audio Source

You can provide audio in two ways:
  • Upload File: Upload an MP3 or WAV file (Max 10MB).
  • Record Audio: Record a 10-second sample directly in the browser.
3

Define Voice Details

  • Voice Name: Give your voice a recognizable name.
  • Gender: Select Male, Female, or Neutral.
  • Description: Optional notes about the voice characteristics.
4

Create

Confirm you have the rights to use the voice and click Create Voice.

​
Using Your Clone Voice on the Dashboard

  1. Go to the Text to Speech page.
  2. Select the Advanced Model (v2).
  3. Open the Voice Selector.
  4. Your created voices will appear under the “Your Voices” or “Custom” section.

​
Using Your Clone Voice in the API

You can use your cloned voice programmatically by referencing its unique voice_id.

​
1. Find Your Voice ID

In the Clone Voices dashboard, find your voice in the “Your Voices” grid. You can copy the Voice ID from there (if available) or use the API to list your voices.

​
2. Make the API Request

Set the voice_id parameter to your Voice ID and the model parameter to advanced. 
import audixa

audixa.set_api_key("YOUR_API_KEY")

audio_url = audixa.tts_and_wait(
    "This is my custom cloned voice.",
    voice_id="custom_voice_12345",
    model="advanced",  # Required for custom voices
    speed=1.0,
)

​
Billing

Clone Voices utilize the Advanced Model.
ModelCost
Advanced4 credits per token

Understanding Credits

See our detailed guide on credit calculation.

​
Managing Your Voices

You can delete voices you no longer need from the Clone Voices dashboard. Note: Deletion is permanent and cannot be undone.