Images to RGB

This is a plumbing node, and plumbing nodes are the ones that quietly save your afternoon. Images to RGB takes an image batch and forces it to plain three-channel RGB - dropping an alpha channel if there's one, converting from grayscale or some other mode if that's what came through. When a downstream node chokes with a shape or channel error, this is often the one-node fix.

Why do you hit this at all? Because not every image that flows through ComfyUI is clean RGB. Load a PNG with transparency and you get RGBA (four channels). Some nodes output single-channel or oddly-shaped tensors. Downstream nodes that assume three channels then throw an error, or worse, silently produce garbage. Images to RGB normalizes everything to the format most nodes actually expect.

How it works

It inspects the incoming image batch and, per the WAS description, converts anything that isn't already RGB - RGBA, grayscale, other modes - into RGB. An alpha channel gets discarded (the transparency is thrown away, the color kept). The result is a standard three-channel image batch that behaves predictably everywhere else in your graph.

Think of it as a "make this normal" adapter. You don't reach for it on purpose so much as slot it in when something two steps later complains about channels.

The inputs and outputs that matter

One input - the images you want normalized - and one output, the same batch as RGB. There's nothing to configure. Drop it in the wire between the node that produces the awkward image and the node that can't stomach it.

That's the whole node. Its value isn't in options, it's in being in the right place.

How to install it

Part of WAS Node Suite. Install once via ComfyUI Manager (search WAS Node Suite, install, restart) or by hand:

cd ComfyUI/custom_nodes
git clone https://github.com/WASasquatch/was-node-suite-comfyui
pip install -r was-node-suite-comfyui/requirements.txt

then restart ComfyUI. Windows portable users should run the pip line with python_embeded\python.exe -s -m pip install -r ... (or the install.bat) so it targets ComfyUI's Python. No models, no heavy deps for this particular node - it's a format converter.

Common issues & troubleshooting

You needed the alpha channel and it's gone. This node strips transparency - that's the trade for getting clean RGB. If a later step depends on the alpha (compositing, masking), pull the alpha out into a mask before you flatten, using a mask/channel node, so you keep the information instead of discarding it.

The error didn't go away. Not every "shape mismatch" is a mode problem. If a node wants a specific resolution and yours doesn't match, RGB conversion won't help - you need a resize. Read the actual error: channel/mode complaints (4 vs 3 channels, "L" vs "RGB") are what this node fixes; size complaints need a resize node.

Colors look flat after conversion. If the source was a mode with baked-in transparency over a dark or odd background, discarding alpha can reveal whatever was underneath. Composite onto a known background (Image Blank makes a solid one) before converting, so the flatten produces the color you expect.

The suite won't load. WAS Node Suite is a large pack that hasn't been maintained since late 2023, and the recurring headache is an "Import Failed" after a ComfyUI update - a dependency version clash, opencv being the usual suspect. Reinstall its requirements against your ComfyUI Python and restart. This particular node is trivial and reliable; it's the suite-wide install that's the weak link.