🌅Image Brightness
Brightness correction without leaving the graph — Image Brightness is one slider and done
- image
- IMAGE
Sometimes a render comes out of the VAE a little too dark and you just want to nudge it without opening an image editor. Image Brightness is exactly that: one image in, one factor, one image out. It's the most boring node in this pack, and that's fine - boring is what you want from a brightness control.
How it works
It's a thin wrapper around PIL's ImageEnhance.Brightness. The tensor is converted to a PIL image, enhance(brightness_factor) is applied, and it's converted back to a torch IMAGE. One input matters:
image- the IMAGE to adjust.brightness_factor- FLOAT, 0.0 to 2.0, step 0.1, default 1.0 (which means no change). Below 1 darkens, above 1 brightens, 0 gives pure black.
Output is a single IMAGE, ready to wire into a Save Image or further post-processing. It's marked as an output node, so the result previews on the canvas, but you can still pipe it onward - the OUTPUT_NODE flag just tells ComfyUI to render a preview, not that the data ends here.
When you'd bother
Correcting a VAE decode that came out muddy, balancing exposure across a few frames before compositing, or lightening a dark img2img pass before it hits a detailer. It's linear brightness, so don't expect gamma or curves - for that you'd be looking at a proper post-processing suite. This is the quick knob.
Where people get burned
The one real gotcha is that the node round-trips through PIL and force-converts to RGB. Feed it an RGBA image and the alpha channel is silently dropped. If you're compositing with transparency, do your brightness before the alpha appears, or use a node that preserves channels. It's also worth knowing the brightness step is 0.1 - fine for subtle work, slightly coarse if you wanted fine granularity, but you can always type a value in.
Honestly? Not much else to go wrong. It's a pass-through with a multiplier. If it "doesn't work," check that the factor isn't sitting at 1.0 (the no-op default) - that's the most common false alarm.
Install
This ships in the ComfyUI Production Nodes Pack by Uriel Deveaud (KoreTeknology). Install via ComfyUI Manager (search "ComfyUI Production Nodes Pack") or manually:
cd ComfyUI/custom_nodes
git clone https://github.com/KoreTeknology/ComfyUI-Nai-Production-Nodes-Pack
cd ComfyUI-Nai-Production-Nodes-Pack
pip install -r requirements.txt
Restart ComfyUI. No models, no downloads - just the usual PIL/opencv requirements. The repo even ships a test workflow for the brightness/contrast pair, which is a nice touch from the author. Find it under image/postprocessing; like the rest of the pack it's filed into existing menu categories rather than a new root folder. Note the class name is Brightness Image (with a space), which is also what appears in workflow JSON.
Inputs (2)
| Name | Type | Default | Description |
|---|---|---|---|
| image | IMAGE | — | |
| brightness_factor | FLOAT | 1.00–2 | — |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| IMAGE | IMAGE | — |