πββ¬ Display String
See any string on the canvas β and force the graph to actually run
- text
- unique_id
Display String is the debugging node you didn't know you needed: wire any string into it and it renders the value right on the node, so you can see mid-graph what a value actually is instead of guessing. Feed it the filename output from this pack's LoRA Metadata node, a processed prompt, or anything else that flows as text, and it shows up in a read-only box on the canvas after the run.
There's a sneaky second benefit. The node is an output node, meaning ComfyUI treats it as an endpoint of the graph. Sticking one at the end of a chain forces ComfyUI to actually execute that chain to produce the value - which is exactly what you want when a workflow "silently" isn't running something and you need eyes on it.
How it works
The node takes a text input, marks itself OUTPUT_NODE = True, and returns two strings. The frontend half does the real trick: a small JS extension (string_display.js) adds a read-only multiline widget to the node and fills it with the executed value. It's deliberately wired so the displayed value is not saved into your workflow JSON - the widget's serialized value is cleared on save, referencing a known ComfyUI issue about widget values leaking into saved graphs. So the display is live proof, not documentation.
The input that matters
text- aSTRINGinput withforceInputenabled. That flag means you can't type a value into the node; it must be wired from something upstream. This trips people up on first use. If you want to just type something, feed it from a plain text or multiline text node.
Outputs
text- passes the value straight through, so the node is safe to inline in a string chain.unique_id- the node's own id as a string, mostly used internally by the display logic. You'll rarely touch it.
Installing it
From ComfyUI Manager, search "Super Loader", or:
cd ComfyUI/custom_nodes
git clone https://github.com/markuryy/ComfyUI-SuperLoader
Restart ComfyUI after either. There are no extra pip dependencies and no model downloads - the pack's pyproject.toml declares an empty dependency list. It's MIT-licensed.
Gotchas
- You can't type into it.
forceInputmeans a wire or nothing. This is the most common "is it broken?" moment. - The shown value doesn't persist in the saved workflow. Reopen the file and the box is empty until the next run. That's intentional, so don't rely on it as embedded documentation.
- The display updates only when the node executes, so if upstream is bypassed or muted, the box stays empty - which is itself useful diagnostic information.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| text | STRING | β |
Outputs (2)
| Name | Type | Description |
|---|---|---|
| text | STRING | β |
| unique_id | STRING | β |