AI voice cloning that sounds like you

You can only clone your own voice or a voice you have the legal right to use. By uploading a voice file, you confirm that: you have permission to use any third-party voice included in your submission; you consent to the use of your uploaded file for AI training to generate your custom voice; your generated voice will remain private and will not be shared publicly on Typecast; you will not use the generated voice for hate speech, false information, defamation, or any illegal or harmful purpose; and you agree to our Terms of Use and acknowledge that the Voice Cloning Policy may change with prior notice in accordance with company policy.

Don't worry! If you didn't succeed, check the [Issue report]. If you encounter any issues, we'll send you an email with instructions to the email address you signed in with. Most likely, the problem occurred due to errors in the sound file. In such cases, we suggest uploading a new voice file by clicking [Generate again] and continuing with the training. To improve your chances of success, try removing any background noise from your voice file, or consider using a different file with consistent sound quality. And if you decide to change your mind, you can delete the custom voice you're creating by tapping the trash can icon next to [Learn More]. (*Please note: Deleted custom voice information cannot be recovered!)

To generate a new Custom Voice, simply tap [Generate again]. This action will erase the current custom voice creation data and take you back to the initial screen. (*Please note: Once a custom voice is deleted, it cannot be recovered.)

If you've created a custom voice and wish to make a new one with a different voice, start by deleting the existing custom voice. (*Please note: Deleted custom voices cannot be recovered). If your subscription hasn't renewed automatically due to an error, etc., deleting an existing custom voice will return you to the Voice Cloning's initial screen.

For effective AI training, it's best to have recording files that are at least 20 minutes long, up to 2 hours.

It's best not to mix too many emotions during training, as it can make it difficult for the AI model to learn effectively, resulting in poor outcomes.

Try reading passages from books you enjoy or articles from the internet to capture a variety of words and sentences. Reading with a consistent emotion and tone is best for effective learning.

Recording in a space with minimal noise is recommended for better quality. Background noises like mouse clicks, wind, coughing, or other voices can interfere with the AI's learning and degrade the results.

You can create custom voices with both the Pro and Business plans. However, if you switch from the Business plan to the Pro plan, only the most recently created custom voice will remain active. The others will no longer be accessible.

The Pro plan includes one custom voice slot. With the Business plan, you get two custom voice slots by default, and you can purchase additional slots if needed.