Interested users can apply for early access on the Character.AI website, and a waitlist is available for those who wish to join later.

Character.AI has officially launched AvatarFX, a revolutionary AI tool designed to transform still images into dynamic motion videos where characters can speak, sing, and express emotions.

This innovative tool stands out by generating high-quality videos of both 2D and 3D characters, including non-human faces, without the need for text-to-image generation.

AvatarFX is versatile, enabling users to animate a wide range of subjects—from human avatars to animals and mythical creatures—using just a single image.

The tool incorporates multiple safety measures to prevent misuse, including filters to avoid creating deepfakes and restrictions on using images of minors and public figures.

Currently, AvatarFX is available only to CAI+ subscribers during its private beta phase, with plans for broader public access in the coming months.

While a specific launch date for the public release has not been announced, the feature is expected to roll out to subscribers soon.

To ensure responsible use, Character.AI has established strict terms of use for AvatarFX, prohibiting impersonation, bullying, and unauthorized use of intellectual property.

AvatarFX employs advanced flow-based diffusion models to maintain temporal consistency, ensuring smooth and realistic movements throughout the videos.

This technology not only synchronizes lip movements and gestures but also enhances the overall realism of the animated performances.

The launch of AvatarFX marks a significant shift for Character.AI, expanding their focus from text and image models to include advanced video generation capabilities.

With over 10 million AI characters available for interaction, Character.AI continues to push the boundaries of digital storytelling and content creation.