To get the best possible voice clone, follow these tips:
Experiment with different audio samples. Don’t just upload once—try different combinations of your audio clips. If your first attempt doesn’t sound right, try again with a new mix, or even create a new clone using the same samples. Results can vary, so don’t be afraid to delete, edit, and repeat until you’re happy with how it sounds.
Limit your total audio to 3 minutes. Adding more than three minutes won’t make your clone better. Focus on clear, high-quality samples of your voice—just you talking, not a dialogue or group conversation. For better accent reduction, consider using around 1 minute of high-quality speech where the speaker’s natural tone is most evident. This minimizes accent-related distortions in the cloned voice.
Bring extra enthusiasm. The biggest improvement comes from how you speak. Avoid robotic delivery by speaking with energy and excitement. A helpful trick: say the same 1 minute block text twice, making your second take even more animated
Imagine your audience. Picture yourself reading to a room full of children or sharing a major scientific breakthrough with a big, general audience. This mindset helps bring out the warmth and excitement that makes a great voice clone.
If you are finding your voice clone is inconsistent (i.e. one speech block sounds different than another), try reverting to our "Classic" model by going to your host and clicking on "Advanced Settings" and adjusting the voice model.
With a bit of experimentation and expressive delivery, you’ll end up with a voice clone that sounds lively, natural, and engaging.