Instant Dream AI released a trailer for the launch of a new feature on its official social media. The video shows that OmniHuman, a new multi-modal video generation model, allows users to generate a vivid AI video by simply entering an image and a piece of audio. According to the information on the OmniHuman technology homepage, the model is a closed-source model developed by ByteDance.

The person in charge of Instant Dream AI said that the Dream Club will set up a strict security review mechanism for this function, and mark the output video with a watermark as a reminder to ensure that AI plays a positive and positive role.