Duix offers two types of AI Avatars: Stock AI Avatars and Personal AI Avatars—both powered by our advanced AI engine. Personal AI Avatars can be customized using either a photo or a video, making it easy to create a lifelike digital version of yourself in minutes.

Discover how to get the best results when creating your AI Avatar and learn what makes a high-quality input essential.

​
Training Personal AI Avatars with Video

To train your avatar using a video, simply upload a 10-second to 1-minute clip that includes at least 5 seconds of clear audio for voice cloning. We’ll guide you to select the most natural 5-second segment to help capture your authentic expressions and voice, ensuring a realistic face-to-face feel.

​
Training Personal AI Avatars with Photo

Prefer using a photo? No problem. Just upload a high-resolution image showing your facial features clearly. For voice cloning, you’ll also need to submit a 5–10 second audio clip separately.

​
How It Works

AI Avatars learn how you look, speak, and move—capturing your voice, facial expressions, and gestures. The result: hyper-realistic videos that look and sound just like you, even in different languages.

The quality of your input—whether it’s a photo or video—plays a major role in the final output. High-quality inputs lead to more lifelike and accurate results.

​
Fast, Flexible, and Shareable

Your Personal AI Avatar takes just 15–20 minutes to train. Once it’s ready, you can reuse it anytime during your subscription period—no need to retrain.

Easily create a custom conversation link to interact with your avatar, and share it with friends, clients, or your audience for an interactive experience. You can also embed your avatar into your own app or website—giving you a talking, expressive digital twin that speaks on your behalf.

Here’s an example of an output from one of our Stock AI Avatars: