Show Anything 🐑

ShowAnything is the debug node. Wire literally anything into it - a string, a number, a seed, a file path, the output of a prompt generator - run the graph, and it prints the value right there on the node face so you can see what an upstream node actually produced. It's print() for a node graph, and once you've used one you stop guessing what's flowing down a wire.

It ships in ComfyUI-MieNodes, MieMieeeee's utility pack, and it's easily the most reached-for node in the set because it's useful in every workflow, not just the LoRA-prep ones the pack was built around.

Why you actually want it

ComfyUI hides intermediate values by default. You build a chain - an LLM rewrites your prompt, a node reformats it, another concatenates a trigger word - and you have no idea what the string looked like at step two unless you go read it somewhere. When output comes out wrong, the usual advice is to isolate the failure by inspecting each stage. ShowAnything is that inspection. Drop it on any wire, run, read the value. It's the fastest way to answer "is the problem upstream or downstream of here."

The pack's own README leans on it for exactly this: in the Kontext prompt workflow it displays the LLM-generated prompt in both English and Chinese so you can eyeball the rewrite before it hits the sampler.

How it works

The input is anything, typed * - the ComfyUI wildcard, meaning it accepts any socket type without complaint. Whatever comes in gets stringified and rendered on the node. There's no configuration; it just shows you the thing.

It's also marked as an output node, which matters more than it sounds. Output nodes force their branch of the graph to execute even if nothing downstream consumes them. So you can hang a ShowAnything off a side branch purely to inspect it, and ComfyUI won't prune that branch as dead - it'll run it to feed the display.

There's one output, a STRING carrying the same value. That means ShowAnything doubles as a passthrough: inspect a value and keep piping it onward in the same spot, no tee needed. Handy when the thing you want to see is mid-chain and you don't want to break the chain to look at it.

Installing it

Grab the whole pack. In ComfyUI Manager, search ComfyUI-MieNodes (ComfyUI_MieNodes), install, restart. Or manually:

cd ComfyUI/custom_nodes
git clone https://github.com/MieMieeeee/ComfyUI-MieNodes

then restart. No model download - it's pure Python. The nodes land under a sheep-emoji 🐑 MieNodes menu.

Common issues

Not much breaks here, because the node barely does anything. Two small things worth knowing:

It stringifies everything. Feed it an IMAGE or a latent and you won't get a picture - you'll get a text representation (a tensor summary, essentially), because the output type is STRING. It's built to show text-ish things: prompts, numbers, paths, logs. For images you want a Preview Image node instead.

You're not short on options. Every mature utility pack has one of these - rgthree's and the Impact Pack's "show text" nodes do the same job - so if you already run those, you don't strictly need this one. But if you've got MieNodes installed for its caption tools anyway, ShowAnything is right there and it accepts the * type, which the stricter text-only variants sometimes don't. It's the one I'd leave on the canvas by default.