What is ControlNet?

How it works

ControlNet provides structural control over AI image and video generation. You provide a conditioning input (edge map, depth map, pose skeleton, or other spatial information) and the model generates content that follows that structure while still being creative with textures, colors, and details. This bridges the gap between the creative freedom of AI generation and the precision control that professional production requires. For example, you can use a depth map from a 3D scene to control the spatial layout, use a pose skeleton to control character positioning, or use an edge map to control the composition. ControlNet is most commonly used with open-source models and self-hosted workflows.