Load Image (Workspaces)

Let's be direct: this node doesn't work yet. "Load Image (Workspaces)" ships in the comfyui_user_workspaces pack as a stub - the load logic is commented out in the source, and the node's function literally returns None. If you found it in a workflow and it's erroring or silently producing nothing, that's not you doing something wrong. It's the node.

I checked the actual code, because the pack README doesn't even mention this node. In py/nodes/nodes.py the class fot_Image advertises two outputs, IMAGE and MASK, but the INPUT_TYPES method contains a typo ("reauired" instead of "required"), so the node registers with no inputs at all. The loading code that would read an image file and produce those outputs is there, fully written, sitting in a comment. The process function that ComfyUI actually calls returns (None,). Whatever the author's plan was - presumably a workspace-aware image loader so you could pull a source image from inside a project folder - it hasn't been finished.

So what do you actually need?

Honestly, nothing from this node. ComfyUI's built-in Load Image already does everything this node wants to do: pick an image from your input folder, get IMAGE and MASK out, ready for img2img, inpainting, IP-Adapter, ControlNet, whatever. The only thing the workspace version would add is loading from a project subfolder, and you can get that today by dropping your source images into ComfyUI/input/ (or a subfolder there) and using the stock loader's folder picker.

If you're building a workspace-organized workflow with this pack, the intended pattern is the other direction anyway: fot_Workspace defines the project, fot_Folder builds a save path, and you save outputs into the project. Loading an input image from inside the project folder is a gap the pack just hasn't filled yet.

Installing it (if you still want to poke at it)

Same install as the rest of the pack. Manager's method is still TODO and the pack can't be published to the ComfyUI registry, so clone it:

cd ComfyUI/custom_nodes
git clone https://github.com/wgedeon/comfyui_user_workspaces

Restart ComfyUI. The only dependency is the tiny python_path pip package. No models, no heavy installs.

The honest verdict

Check back in a few releases - the image-loading code is already written in the repo, just commented out, so the author may finish it at any point. If a future update activates it, it'll probably behave like Load Image with a folder scoped to the current workspace, which would actually be handy. Until then: use core Load Image, wire fot_Folder's path_full into your save node, and enjoy the one part of this pack that does work. This is a young, low-adoption project (last commit is literally "debugging registration"), so expect rough edges like this one.