Nourivex Load Image

Nourivex Load Image is ComfyUI's built-in LoadImage, wrapped so it lives in the Nourivex namespace. It does exactly what the core node does - you pick an image from your input folder (or upload one), and you get an IMAGE tensor out - and the only real reason it exists is so a Nour_Comfy workflow can be assembled entirely from pack nodes, which is the author's design philosophy across this whole I/O family. If you're running the pack anyway, use it; if you're not, the core node is identical.

What you get

• image - the dropdown listing every image in ComfyUI/input, with an upload control attached. Same widget, same behavior as core LoadImage.
• Outputs: IMAGE and MASK.

The MASK output is the one worth a moment. It's only populated when the loaded image has an alpha channel - or when you've dropped in a masked image. Connect it when you're feeding an inpainting graph and you actually want the mask to ride along; leave it disconnected otherwise. A lot of beginners wire it everywhere reflexively, which is harmless, but it's only doing work in those two cases.

The one documented failure mode

This node carries the pack's most specific troubleshooting entry, and it's worth knowing because it looks scary and isn't. If you see:

Nourivex Load Image requires ComfyUI runtime

it means the wrapper failed to resolve the core LoadImage node at import time. The wrapper tries a few module locations (nodes, then the comfy_extras fallbacks) to find it, and when all of them fail, it degrades gracefully to a plain text field instead of a working loader. The README's fix: update the pack, then restart ComfyUI. In practice this bites after a ComfyUI update shuffles its internals, and a restart after the pack is refreshed is the cure. If it keeps happening, check the console log to make sure the Nour_Comfy folder actually loaded.

Install

One pack, one install:

cd ComfyUI/custom_nodes
git clone https://github.com/Nourivex/Nour_Comfy

Restart ComfyUI - or use ComfyUI Manager and search for Nour_Comfy. No dependencies, no downloads. Honestly? If this is the only node you need, don't install the pack for it; the core LoadImage you already have does the job. Grab this one because the pack's workflows reference it and you want those to run without missing-node errors.