Rebel Load Image 📂

Every img2img, inpainting, or image-reference workflow starts the same way: get a picture into the graph. Rebel Load Image 📂 is the pack's version of ComfyUI's LoadImage node - pick a file from your input folder and it hands you the IMAGE tensor plus an optional MASK, with the pack's animations playing behind the file picker. It's a straight drop-in for the stock loader: same inputs, same outputs, same behavior, fancier box. If you've been using stock LoadImage to feed VAE Encode or an image-to-video model, swap this in and nothing else changes.

How it works

Standard wrapper: pull the native LoadImage inputs, add theme and color_scheme, delegate the file read to core ComfyUI. The image decoding, color handling, and mask extraction are all ComfyUI's own code - the animated background is a frontend overlay with zero effect on the loaded pixels. When ComfyUI updates its loader, this node inherits the update automatically because it just calls through.

The inputs that matter

• image - the file to load, from the input folder dropdown. That folder is where ComfyUI expects your source images; drop files there and restart (or rescan) so they show up.
• theme / color_scheme - the 12 animations and 14 palettes. Decoration.

Outputs are the classic pair:

• IMAGE - the loaded image tensor, feeding your VAE Encode, ControlNet preprocessors, or anywhere pixels are needed.
• MASK - the image's alpha channel as a mask, which is how transparent PNGs carry an automatic mask into inpainting-style workflows. If the image has no alpha, you get an empty mask.

Installing it

One clone, no dependencies, no model downloads:

cd ComfyUI/custom_nodes
git clone https://github.com/RealRebelAI/Rebels_Animated_Nodes.git

Restart ComfyUI, then hard refresh the browser (Ctrl + F5 / Cmd + Shift + R) - the pack's documented critical step, since ComfyUI caches frontend JS and the animation won't draw until you force a reload. It's under Rebel AI → Animated Core.

Where people get burned

• "My image isn't in the dropdown." It reads the input folder only. Put the file there and restart ComfyUI so the folder re-scans - same as with the stock node.
• "The MASK output does nothing." It only carries a mask if your image has an alpha channel. A plain JPG or opaque PNG gives you an empty mask; that's expected, not a bug.
• Animation missing after install. Browser cache - hard refresh, and re-refresh after updating the pack.
• Flat backdrop on Nodes 2.0. The pack renders on the legacy LiteGraph canvas; the image loads fine but the animation can go missing under ComfyUI's newer frontend.

A simple, faithful reskin of a node you'll use constantly if you do any image-in work. The MASK passthrough is worth remembering - it's the quiet reason this loader is more useful than it looks for inpainting.