Populate the Scene
Add realistic characters matching the scene perspective.
- Free to try
- No plugin
- No credit card
Nano Banana Pro
Sample input · click Generate to see the result
About this model
Adds realistic human figures to an architectural render or photograph at a controllable density. Figures are scaled to the scene perspective, clothed in muted contemporary wear, posed in plausible activities (walking, talking, sitting, looking), and positioned to give scale to the architecture without obscuring critical views. Shadows are cast consistently with the existing sun direction; ground contact reads correctly. Density is adjustable: from one or two figures for a residential courtyard to a small crowd for an urban plaza. Output preserves geometry, camera and lighting exactly. Useful for competition boards, presentation decks, and any visual where the absence of human scale makes the architecture feel inert. An empty 3D render reads as a model; the same render with three plausible figures reads as a place.
When to use it
A landscape architect submits a competition for a 4,500 m² public square redesign. The 3D renders are technically beautiful but read as cold — the empty plaza looks like a 3D model, not a place. The architect runs each hero shot through Populate the Scene at medium density, gets a small crowd of plausible figures walking, sitting on benches, gathered near the fountain. The competition jury sees a place, not a model.
Not at the moment — the density selector controls how many people; their activity is inferred from the scene type (people walking on a sidewalk, sitting at a café terrace, gathered near a building entrance). The model picks plausible behaviour for the architectural context.
The model defaults to a varied contemporary mix appropriate to the scene type — adults of mixed ages and genders, occasional children and elderly figures. It does not produce ethnic-specific or demographic-targeted casts.
