⚡ Quick Answer
ControlNet is a model that adds extra, structural guidance to image generation — a pose, a depth layout, or an edge outline pulled from a reference image — so the output follows that structure while your text prompt and checkpoint still handle everything else, like style and detail.
A regular prompt is loose about composition — it can't reliably force a specific pose or camera angle. ControlNet fixes that by feeding the sampler a literal structural map to follow, on top of the usual text-based guidance.
Where You'll See It
Load ControlNet Model loads a ControlNet file, and Apply ControlNet takes that model plus your positive and negative conditioning and a preprocessed reference image, then outputs new conditioning that feeds into the KSampler.
Quick Example
A Canny Edge preprocessor turns a reference photo into a simple line-art outline. Connect that through a ControlNet and the sampler keeps the same composition and edges from the photo, while your prompt and checkpoint completely change the style — turning a photo into a painting, for example, without losing the original layout.
Frequently Asked Questions
See It In Action
Ready to control your composition?
Our ControlNet guide covers preprocessors, model types, and strength tuning with real examples.
Published: 2026-08-25 · Last updated: 2026-08-25
Join the discussion
Sign in to leave a comment or reply
No comments yet
Be the first to share your thoughts!
