Earngenix Logo
Skip to main content

Glossary · ComfyUI Basics

What Is ControlNet in ComfyUI?

By Earngenix Team ·

⚡ 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.

Most ControlNet models expect a specific preprocessed input type — depth maps for depth ControlNets, pose skeletons for OpenPose ControlNets. Using the wrong preprocessor output for a given model usually produces broken or ignored guidance.

Frequently Asked Questions

Almost always, unless you already have a matching depth map, pose skeleton, or edge map ready. A preprocessor node (like Canny Edge or OpenPose) converts a normal photo into the guide format the ControlNet model expects.

It sets how strongly the reference structure is enforced. Higher strength locks the output closer to the reference composition; lower strength lets the checkpoint and prompt have more influence over the final layout.

Yes — chain multiple Apply ControlNet nodes together, each with its own model and reference image, to combine guidance types like pose and depth in a single generation.

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

Discussion

Join the discussion

Sign in to leave a comment or reply

💬

No comments yet

Be the first to share your thoughts!