LayerColor: AutoAdjust

AutoAdjust is the "fix my image automatically, then let me tweak it" node. It's basically Photoshop's Auto Levels/Auto Color idea: it looks at the image, stretches the tonal range and balances the color for you, and then hands you a set of manual knobs to push the result in whatever direction you want. When a render comes out flat, dull, or slightly off-color, this is the one-node cleanup that gets you 80% of the way there before you touch anything.

The mechanism is auto-correction first, manual override second. The automatic pass does the histogram work - stretching contrast, neutralizing casts - and then your slider values get layered on top. The inputs that matter:

• strength - 0 to 100, default 100. How hard the automatic correction is applied. Drop it if the auto pass is too aggressive and you want a gentler version.
• brightness / contrast / saturation - each -100 to 100, default 0. Manual adjustments applied after the auto pass. This is where you dial the final look.
• red / green / blue - each -100 to 100, default 0. Per-channel color nudges on top of the auto color-balance, for when the automatic neutralization didn't go quite where you wanted.

Output is a single image.

The reason to reach for AutoAdjust over the individual LayerColor nodes is convenience: instead of chaining Exposure + Gamma + ColorBalance + saturation, you get an automatic first pass and a full set of manual controls in one box. For batch work especially - running a pile of images through the same cleanup - the auto pass adapts per image, so a flat one gets more correction than an already-good one, which a fixed manual chain can't do.

The trade-off is the usual auto-correction trade-off: it's guessing. On a well-exposed image the auto pass can over-correct, pushing contrast or shifting color you didn't want moved - that's exactly what the strength slider is for, so turn it down (or to 0, making the node a pure manual adjuster) when the automatic behavior fights you. It also has no concept of intent: a deliberately low-contrast, moody frame will get "fixed" into something punchier, which may be the opposite of what you were going for. Treat it as a fast starting point, not a final grade, and it's genuinely handy.

Install is the pack standard. Easiest is ComfyUI Manager - search ComfyUI Layer Style (or "LayerStyle"), install, restart. Manual: cd ComfyUI/custom_nodes && git clone https://github.com/chflame163/ComfyUI_LayerStyle, then pip install -r requirements.txt from inside the folder using ComfyUI's Python, and restart. On a hosted ComfyUI it's usually already in the image.

The pack-wide caveat: Layer Style is notorious for import failures - install it, restart, and the whole node set is red or missing. It's rarely the individual node; it's a dependency clash. The pack pulls a heavy stack (transformers, onnxruntime, opencv), and one mismatched version - a stray transformers or tensorflow in the environment being the classic - takes the entire pack down on import. The fix: open the ComfyUI startup log, find the traceback (it names the package that actually failed), reinstall requirements against ComfyUI's own Python, and use Manager's Try Fix. If AutoAdjust is missing, every LayerStyle node is missing, and that shared import failure is what you're really solving.