Interior AI developer Pieter Levels announces a new feature: editable 3D scenes generated from videos, photos, or floor plans
Independent developer Pieter Levels (@levelsio) posted on X that his app Interior AI is adding a new feature for AI-generated headless Blender 3D scenes. Users can upload a rough room video, several photos, or a floor plan to generate an editable 3D scene.
Previously, the app mainly relied on image models, taking room photos as input and outputting redesigned results. The new feature supports arbitrarily changing wall, floor, and ceiling colors in the scene, and moving, deleting, or adding furniture from a large 3D furniture library. Levelsio said he has already made it possible to upload a video, photo, or floor plan and then clear the room, adjust furniture, apply paint, and perform other operations, and plans to add a top-down floor plan view, one-click switching between interior design styles, and full integration into Interior AI.
This feature is based on recent advances in AI’s ability to generate 3D scenes from limited inputs. Levelsio said floor plans were previously difficult for image models to understand. He said he is working to launch the related feature within hours.
Original source: x.com/levelsio/status/2120639435318464513
ABAB AI Insight
An independent app developer has integrated AI-generated 3D scene capabilities into an interior design tool, enabling the transformation from limited input to editable models. levelsio announced that Interior AI has added a feature to generate Blender 3D scenes from video, photos, or floor plans, allowing users to directly modify wall and floor colors and furniture positions. The app previously focused on image-to-image redesign, and the new path introduces editable three-dimensional structures.
Unlike static image output, this path turns the generated result into an interactive scene where users can clear, move, and recolor objects. In similar cases, capital is more concerned with generation stability and the size of the furniture library, as well as whether it can reduce the upfront visualization cost of interior design. Floor plan inputs were previously difficult for pure image models to handle, and they have now become a new capability.
This update expands Interior AI from image redesign to editable 3D scenes. Beneficiaries are users seeking rapid visual modifications; the developer gains a new feature iteration path. The structural change is that the interior design tool is shifting from image output to AI-generated editable 3D models, and the input types are expanding from photos to video and floor plans.
Cognitive law
AI generating editable 3D scenes from limited input turns static redesign into interactive modification.
Floor plans and video become new inputs, expanding the scope of understanding beyond the previous image model.
An independent developer rapidly iterates, directly integrating emerging 3D generation capabilities into an existing app.