Skip to main content
This feature is available to Business and higher-tier members.

​
What Is a Variant

A variant is a new output generated from the same Talking Photo project using a different audio.
It allows you to produce multiple versions while keeping the same trained model.

​
Why Create a Variant Instead of a New Project

Creating a variant reuses the model that was already trained in your project.
Compared with starting a new project:
  • ⚡ Much faster — usually completes within minutes instead of tens of minutes or hours.
  • 🎯 More consistent — keeps the same facial expressions and movements for a stable visual quality.  
  • 💰 Cost-saving — only 5 points per minute are charged, with no additional 5 base points.

​
How to Create a Variant

1

Open Your Talking Photo Project

Go to your Talking Photo project and click Create with New Audio in the top-right corner.
2

Upload or Input New Audio

Upload a new audio file, or enter text to generate speech via Text to Speech.
3

Generate

Click Generate and wait for the process to complete.
4

View Variants

All variants are stored under the same project.
Click the Variant option in the top-right corner to open the dropdown and switch between variants, preview, or download.