Node Search Utility

Every ComfyUI user has had the same frustration: you remember a node's name but can't find it in the giant right-click menu, or worse - you're reading a workflow's JSON and you need to know the internal class_name that a display name maps to. Node Search Utility is a tiny search box for your own installed nodes: type a query, pick whether you're matching class name, display name, or category, and get back a list of matches plus a count. It's a tool for navigating your own install.

It's part of DebugPadawan's ComfyUI Essentials, a small MIT-licensed utility pack. Its target audience is a bit more developer-flavored than the rest of the pack, but the "what is the internal name of this node" question comes up more than you'd think.

How it works

At execution, the node imports ComfyUI's internal nodes module and iterates its NODE_CLASS_MAPPINGS (and the display-name mapping), reading each node's category off its class. It does a case-insensitive substring match against whichever field you chose:

• class_name - match the internal class string (e.g. KSampler).
• display_name - match the human label you see in the menu.
• category - match the menu category path (e.g. latent/upscale).

Leave search_query empty and it lists everything, which is actually useful for inventorying a fresh install.

The outputs and the gotchas

• node_list - a LIST of strings, each formatted as class_name | display_name | category. Note the | separators: if you want the raw class name, you're parsing these strings.
• count_summary - a STRING like Found 3 nodes matching 'sampler'. Human-readable, not a number - don't expect to feed it into an INT input without converting.

The bigger gotcha: it depends on ComfyUI's internal nodes module. If that import fails (say, a deeply modified frontend), you don't get a crash - you get an empty list and a count_summary that says it couldn't access the mappings. That message is the error; there's no red popup.

Install

Standard custom-node fare:

cd ComfyUI/custom_nodes
git clone https://github.com/DebugPadawan/DebugPadawans-ComfyUI-Essentials.git

Restart ComfyUI, or install via ComfyUI Manager by searching "DebugPadawan". requirements.txt lists numpy, torch, and opencv-python - numpy and torch already ship with ComfyUI and the shipped code never imports cv2, so no heavy downloads, no model files, no API keys.

When it earns its keep

• Hunting class names for API calls or workflow JSON edits - the class_name | display_name | category lines are exactly the mapping you were about to grep for by hand.
• Checking whether a node is actually installed before you blame a missing-node error on a workflow.
• Browsing categories - empty query + category mode gives you a readable inventory of your install.

It's a diagnostic, not a daily driver - but the day you need a node's internal name, this beats manually searching GitHub.