ControlNet
ControlNet is a technique that forces an AI image model to follow the structure of a reference — edges, depth or pose — while changing style, materials or content. The geometry stays; everything else can move.
A plain prompt describes what you want but not where things go. ControlNet adds a second input — an edge map, a depth map, a stick-figure pose extracted from a reference image — and the model is conditioned on both. Words pick the style; the map pins the structure.
A concrete example: restyle a photo of your living room as a Japandi interior. Without structure guidance the model moves the windows and invents a new layout; with it, every wall, door and sightline stays exactly where the architect put it, and only materials, furniture and light change.
How Hermosso uses it: the same structure-guidance idea runs through The Studio, where restyle keeps a room's architecture pixel-identical, and AI Relight, which changes a photo's lighting while the subject's identity stays locked.
Try it on your own photos
Upload a few selfies, and Hermosso trains a private AI model of you — then generates studio-quality photos in any style. Your first credits are free.
Create your AI photos →