🔶 controlnet Apply + Streng Start End

Same job as chaosaiart_ControlNetApply - apply a ControlNet's spatial guidance to your positive and negative conditioning - but this version ships with actual slider bounds and sane defaults, where the plain "Apply" node leaves those blank. Strength defaults to 1.0 with a 0–3 range, start defaults to 0.0, end defaults to 1.0, both 0–1. If you're not sure which of the two ControlNetApply variants to grab, this is the one that won't leave you guessing at reasonable starting values.

It's still the "advanced" style ControlNet apply - strength plus a start/end window over the sampling steps - not the plain single-strength version some other packs ship. That window is the parameter worth actually using: ending the condition early (say, 0.5–0.6 instead of the full 1.0) lets the model settle composition from your control image, then finish detail on its own, which is the standing community advice for anything structure-heavy like Canny or pose.

Inputs and outputs

• positive, negative - required CONDITIONING.
• control_net - required CONTROL_NET.
• image - required IMAGE, your preprocessed control image.
• strength - required FLOAT, default 1, range 0–3, step 0.01. This is the one you'll adjust most: too high and the output gets rigid and waxy, too low and the ControlNet barely does anything.
• start - default 0, range 0–1. When guidance begins, as a fraction of total steps.
• end - default 1, range 0–1. When guidance stops. Bring this down if the model's being over-constrained late in sampling.
• Outputs: POSITVE, NEGATIVE - wire into your KSampler.

Installing it

ComfyUI Manager: Install via Git URL → https://github.com/chaosaiart/Chaosaiart-Nodes - the exact method the README documents. A title search for "Chaosaiart" should also find it.

Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/chaosaiart/Chaosaiart-Nodes

Linux (venv activated first, if applicable):

pip install opencv-python
pip install tqdm

Windows: use the bundled Install_windows script. Restart ComfyUI afterward. No extra downloads for this node specifically, but you need a ControlNet model that actually matches your checkpoint's architecture.

Common issues

Architecture mismatch is the number-one failure. SD 1.5-era ControlNet files won't load against SDXL, Flux, or other DiT-based checkpoints - every modern architecture needs its own ControlNet trained from scratch, not converted. Check that your ControlNet's product page or filename actually names your base model.

Some classic condition types don't exist anymore for newer bases. If you're on Flux, Qwen-Image, or another 2025-26 model and looking for something like the old QR-code/brightness ControlNet, it likely hasn't been rebuilt - new-base ControlNets tend to cover canny, depth, pose, and an inpaint mode, and stop there. Stick to those on modern checkpoints.

Composition looks locked in and lifeless. Strength too high combined with end running to 1.0. Drop strength toward 0.6–0.8 and end the condition around 0.5–0.6 through sampling instead of the full run.

Whole-pack import failure. All Chaosaiart nodes share one Python module - a failed opencv-python/tqdm install takes every node in the pack offline, not just this one. Check your ComfyUI startup log for an import error first.