Vidu has introduced Vidu S1—an innovative model for interactive video streaming that allows controlling digital avatars using voice commands in real time.
What Happened
Developed by Vidu, the Vidu S1 model provides 540p resolution at 25–42 FPS on standard GPU hardware. The system supports an infinite video stream and allows for the creation of both realistic characters and anime avatars, integrating voice cloning features to create a closed-loop interaction.
Context
This technology shifts video generation from the classic "prompt-response" mode to an interactive streaming paradigm. This allows digital characters to transform from static objects of observation into active communication participants capable of responding to live speech.
Why It Matters for the Industry
For the industry, this signifies a transition from batch content generation to the creation of live digital avatars for video calls, customer support, and virtual assistants. The availability of an accessible API allows for rapid prototyping of solutions to integrate AI assistants into existing communication platforms such as Zoom or Teams.
Why It Matters for Users
Regular users gain the ability to interact with AI characters as live interlocutors in a video chat format. Controlling an avatar's actions and appearance through voice makes the use of digital human technologies significantly more natural and accessible.
What Is Not Yet Known / Limitations
Moving toward industrial-scale deployment will require addressing challenges in latency optimization and scaling GPU inference. Additionally, there are critical risks related to the use of biometrics and copyright that require further legal regulation.
Sources
Author
Look at AI, Editorial Team