Something Show

Somewhere in a workflow, a value you can't see is driving everything. SomethingShow is the node for that feeling. It accepts literally anything - a number, a string, a tensor, a list, a dict - and shows it as text in the UI, prints it to the console, and, in the genuinely useful bit, writes the value into the workflow data that ComfyUI embeds in your saved PNG. The name is doing what it says: something goes in, it gets shown. It's one of the debug-oriented "show" nodes in Cyber-BlackCat's ComfyUI-MoneyMaker pack (the "ComfyUI-Yuan" one, where yuan is a currency pun and the README promises the nodes will make you real money).

How it works

It's a wildcard-typed output node in the "MoneyMaker😺/show" category, so it will accept any wire you throw at it. Strings pass through as-is; tensors are reported by their shape rather than their contents; anything else gets JSON-serialized or stringified. The result renders in the node's widgets and, via the workflow metadata ComfyUI already stuffs into saved images, gets embedded right alongside the PNG. And despite being marked as an output node, it returns its input unchanged (output = whatever came in), so you can leave it sitting inline in the middle of a graph without breaking the data flow.

When you'd reach for it

Debugging, mostly. Want to know what a node is actually emitting - the seed, a CFG value, the shape of a tensor - without hunting through a wall of console output? Wire it in and it's right there on the canvas. Want a saved workflow that remembers a number computed mid-run, like which candidate won a comparison or a measured metric? This stamps it into the PNG metadata, so the value rides along with the image instead of being lost when you close the session. That beats screenshotting, and it beats trying to remember.

What it is not

It is not an image preview and it is not a saver. Feed it a tensor and you get a shape, not pixels - that's PreviewImage's job. It's not a static note node either, because those don't actually run; this one evaluates and captures real values every execution. The console spam is by design - every run prints the value - which is great while you're debugging and mildly annoying when you're not. One more thing to know: because the input is optional and wildcard, it's easy to wire it in backwards or forget which output you're watching; the shape reporting helps, but it will not tell you much about tensor contents.

Honest take

It's a small utility from an obscure pack, and you could argue ComfyUI's built-in preview and note nodes cover most of this ground. But the "embed a live computed value into the saved PNG" behavior is genuinely handy, and it costs nothing to have installed. If you're the kind of person who stares at an output image wondering which seed produced it, this is a nice belt-and-suspenders addition to your workflow.

Install

Same as the rest of the MoneyMaker pack. ComfyUI Manager - search "MoneyMaker" (pack title "ComfyUI-Yuan") - or by hand:

cd ComfyUI/custom_nodes
git clone https://github.com/Cyber-BlackCat/ComfyUI-MoneyMaker

Then restart ComfyUI. Dependencies are torch, torchvision, numpy, pillow, opencv-python, and scikit-image, with scikit-image the one most likely missing from an existing install. No models to download - this pack is all in-graph math and debugging helpers.