BMAB Preview Text

The text equivalent of Preview Image - drop it mid-graph anywhere a string is flowing, and it shows you the value in the ComfyUI UI without ending the chain. That last part is the useful bit: unlike a pure debug/print node, this one also has an output, so you can inspect a value and keep using it downstream in the same run, rather than choosing between "see it" and "use it."

Where it actually earns its place

The obvious pairing in this pack is right after BMAB Google Gemini Prompt - you send text off to an LLM, get a generated prompt back, and before you trust it to drive a whole generation, you probably want to actually read what came back. LLM output can be verbose, include unwanted preamble ("Here's an expanded prompt:"), or just miss the mark on what you asked for, and none of that is visible if the string goes straight into a CLIP Text Encode with no stop along the way. This node is that stop.

More generally, it's worth scattering through any graph where a node produces or transforms text - a caption, a tag string, a prompt built from concatenated pieces - any point where "is this actually what I think it is" is a question worth two seconds to answer before it feeds a sampler.

The inputs and outputs that matter

• text (required, STRING) - the value to display. Not flagged multiline in the schema, so treat it as a single-line preview rather than a full text editor.
• Output: STRING - the same value, passed through unchanged, so you can keep wiring your graph forward from here.

is_output_node: true in the schema, meaning it also registers as a terminal display point in ComfyUI's execution - you'll see its content surface in the node UI when the graph runs, the same way Preview Image does.

How to install it

ComfyUI Manager: search comfyui_bmab. Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/portu-sim/comfyui_bmab.git
cd comfyui_bmab
pip install -r requirements.txt

No model, no companion pack dependency - this is a display utility.

Common issues & troubleshooting

Nothing shows up in the node when you run the graph. Make sure the node actually executed - if it's off the active path (nothing downstream needs its output and ComfyUI pruned it), some setups won't run a node purely for its side-effect display. Wire its output somewhere, even a dead-end, if you need the preview to reliably fire.

Long text gets cut off or is hard to read in the small node box. This is a UI-space limitation, not a data problem - the underlying string is intact and passed through in full via the output; it's only the on-canvas preview that's cramped. Resize the node wider, or check the value downstream if you need to read the whole thing comfortably.

You're using this to debug and it's slowing down iteration. For a quick one-off check, that's fine, but if you're leaving several of these scattered through a production workflow permanently, consider whether you actually need the visibility every run or just while you were building the graph - pull them out once you trust the pipeline.