Load Image Path (Yogurt Nodes)

Core ComfyUI's LoadImage makes you put files in one specific folder and pick them from a dropdown. That's fine until the image you need is a project render on a different drive, or you want a batch of images driven by a list of paths instead of a hand-picked dropdown. YogurtLoadImagePath is the path-string version: give it an absolute path and it hands back the image, no copying required.

It's part of ComfyUI-YogurtNodes, yogurt7771's 150+ node all-in-one pack. The pack is obscure - MIT-licensed, auto-generated README, no community presence - but the path-based loaders are the quiet workhorses of its IO section, especially once you start driving workflows programmatically.

How it works

A single required input, image_path - a STRING with the path to the image file. Two outputs, identical to the dropdown loader:

• IMAGE - the image tensor for anything downstream.
• MASK - the alpha channel as a mask (transparent pixels become mask values).

That's the whole surface. No dropdown, no folder restriction - whatever path you type is what gets read.

Where you'll use it

• Images stored outside ComfyUI/input, referenced in place.
• Batch processing: build a list of paths with a glob node (the pack's YogurtGlobFiles), then load each path in a loop instead of hand-selecting.
• Workflows where a path arrives dynamically - from a text node, an LLM output, or a config file.
• Paired with path-operator nodes for join/relative-path logic.

The trade-off is the usual path discipline: absolute paths, files existing at queue time, and separator sanity (forward slashes dodge most Windows grief). A missing file errors the graph, so if your path comes from automation, validate it upstream.

One tip if you're mixing loaders: the dropdown loader (YogurtLoadImage) is friendlier for one-off images; this one shines the moment you need a path as data. It's also the natural complement to the pack's SaveImageBridge, which writes back to your output directory - load by path, process, save by path, all without touching the input folder.

Install

ComfyUI Manager, search ComfyUI-YogurtNodes, install, restart. Or:

cd ComfyUI/custom_nodes
git clone https://github.com/yogurt7771/ComfyUI-YogurtNodes
cd ComfyUI-YogurtNodes
pip install -r requirements.txt

Restart and it's under "Yogurt Nodes". Light dependencies (numpy, pillow, requests, opencv-python), no model downloads - path loaders are thin wrappers and install accordingly.