We use the latest state-of-the-art image and video models, and we run them locally for the best speed and security. This lets us integrate them efficiently into our pipeline while keeping full artistic ownership of the final result.
AI supports our process by simplifying the most time-consuming tasks, without touching our creative decisions.
We have been experimenting with WAN 2.2 inside ComfyUI to refine our 3D renders, and the results are very promising. The model adds extra natural detail while preserving our original artistic intent.
When we reach 90 percent realism in a render, the remaining 10 percent usually requires a huge amount of extra time in shading, rendering and compositing.
Using open-source diffusion models as a creative finishing step helps us focus on what matters most: art direction, movement, lighting and mood.
It reduces the technical grind while keeping our signature look fully intact.
During pitches and early project phases, we can use AI image and video models to explore ideas much faster.
We can quickly validate art direction, layouts, lighting and composition, and we can iterate visually until we find the right mood.
This helps our clients see the creative potential early in the process without waiting for full 3D production.
Once we start our real 3D scenes, the prototypes give us a strong base for colors, framing and atmospheres. It accelerates the creative phase while keeping everything under our artistic control.
This ensures perfect sharpness on text, labels and materials at any resolution.
AI is used only to enhance the background and supporting elements, giving us richer details and more dynamic visuals.
We often need to deliver the same animation in several formats. A shot created in a 16:9 landscape ratio might also be required in 9:16 for social media, or in a wider format for in-store screens.
Re-rendering the entire 3D scene each time can be costly, especially when the client requests additional versions late in production.
AI outpainting lets us extend or reframe our shots intelligently. We keep the original render as-is, and AI fills the extra space around it while matching the style, lighting and textures of the shot.
This gives us the flexibility to create new formats quickly, without compromising quality or scheduling a full new render.
