For our latest portfolio project, Coastal Residence, we developed a workflow that combines 3D previs in Blender with AI image generation to give us much more control over the final result.
We first built the entire property as a rough 3D block-out in Blender.
The models themselves don’t need to be beautiful.
They’re there to establish the important things: architecture, scale, room layouts, furniture placement, camera position, lens, composition, and perspective.
Once the shots are locked, those rough Blender frames become the foundation for the final AI-generated images.
Blender controls the structure. AI creates the finish.
This gives our team the ability to art direct each shot before generating the final visual, while keeping the property much more consistent from image to image.
And because everything begins with the same 3D environment, that one model can become much more than a set of still images.
For Coastal Residence, we used the workflow to create photoreal renders, floor plans, cinematic video, property walkthroughs, and AI agent-led tours.
We’re attaching a before-and-after video showing the transformation from the original Blender block-outs to the finished visuals.
Realistic 3D floor plans created from a 2D architectural plan, with detailed furniture, materials, textures, and interior elements based on the client’s requirements and references. Designed for clear presentation and real estate marketing.
Is AI finally good enough to shoot a product demo on its own, or human touch is still needed?
I generated a 20s product demo with Seedance 2.5 in a single take providing it with one product image reference and an example video transcript (posted in the comments).
Do you disclose...