ScenesAi

34B‑Activated 4B Real‑Time Generative Interactive World.
2026 NVIDIA SHOWCASE PROJECT.
NVIDIA INCEPTION PROGRAM.

[ScenesAi @ScenesAi]

A controllable diffusion solution based on a 3-billion-parameter control model, unaffected by camera movement, shake, or occlusion, enabling one-click image rendering and compositing based on webcam input.

[CYANPUPPETS @cyanpuppets]

Real-time video generation technology based on ordinary webcams enables naked-eye synthesis of sci-fi films. Powered by real-time computation on the NVIDIA RTX 5090, it leverages spatial computing and real-time diffusion generation to form a new large-model inference technology.

[CYANPUPPETS @cyanpuppets]

With just a computer, you can generate a world in real time. Explore the world reimagined by AI, summon monsters to battle with your friends, and bring the AI world into your camera. The Dawa Real-Time World Model is rediscovering the laws governing how the world works.

[ScenesAi @ScenesAi]

【Dawa Real-Time World Model】Moving beyond frame-by-frame generation toward persistent world-state maintenance, Dawa explores a generative world model with 34B total parameters and 4B active parameters, designed for continuous operation, real-time interaction.

[ScenesAi @ScenesAi]

A webcam image synthesis solution based on a 3-billion-parameter diffusion control model, enabling one-click generation of high-fidelity 3D virtual scenes.

[ScenesAi @ScenesAi]

@prashantraval85 The software runs locally and uses the computing power of the computer graphics card.

[ScenesAi @ScenesAi]

A controllable diffusion solution based on a 3-billion-parameter control model, unaffected by camera movement, shake, or occlusion, enabling one-click image rendering and compositing based on webcam input.

[ScenesAi @ScenesAi]

@0xStacking We expect to release this model after June.

[ScenesAi @ScenesAi]

@Anweshac1211 Yes, we trained a small model with 3 billion parameters to control diffusion interference.

[ScenesAi @ScenesAi]

@charliiai Motion capture+video generation model.

[ScenesAi @ScenesAi]

A revolutionary AI virtual production solution featuring controllable diffusion and motion capture interaction. AI generates scenes and visual effect assets, while a 3-billion-parameter control model enables pixel-level diffusion control on 2D footage with editable.

[ScenesAi @ScenesAi]

AI one-click generation of 60-second coherent animations that follow real-world physics.

[ScenesAi @ScenesAi]

A new technological study on real-time AI content generation based on 2D camera footage, enabling live creation of indoor snowfall, flowing water, and grassy environments, with performers able to interact with the generated content in real time.

[CYANPUPPETS @cyanpuppets]

The latest free version of UE5 AVATAR is now available! This update enhances real-time rendering quality and adds hundreds of new digital clothing and accessories. Create personalized 3D models, customize outfits, select scenes, and connect AI motion capture for real-time rendered 3D animations. Anyone can create 3D character models—no professional modeling skills required. Future versions will support 3D model import/export and image-based 3D model generation. Get the download link in the comments!

[ScenesAi @ScenesAi]

We will be releasing our first version of the demo in June, which will be the AI tool that will greatly enhance the efficiency of movie making, available on every individual user's computer, phone.