QWEN 文本展示

Every pack that calls an LLM eventually runs into the same wall: the model's answer is a plain text string, and ComfyUI doesn't exactly make plain text strings fun to look at. That's what this node is for. It's an output node that displays whatever text you feed it - right there on the canvas - and passes it through so you can also use it downstream. It's the "show me what the model said" node.

If you're using this pack, you'll reach for it constantly. QWEN3_Text returns a response string, QWEN3VL_Image and QWEN3VL_Video return description strings, and without a display node you're stuck reading model output in the console or guessing what a thumbnail's caption actually was. Drop a QWEN_TextDisplay after any of them and the answer appears in a read-only text box on the node itself.

How it works

Two pieces cooperate. The Python side is marked as an output node (OUTPUT_NODE = True) and accepts a list of strings, so it happily eats batches - feed it the list output from a folder loader pipeline and it'll show every result. The front-end side is a small JavaScript extension that renders each incoming string as a read-only multiline widget on the node. When a new result arrives it clears the old widgets and repopulates, so the node grows to fit whatever you ran last.

The output is a STRING list, which means it's not a dead end: wire it onward into the pack's text-processing nodes, or into a text-encoder, and the display is just a side effect you can also consume. Input and output are the same text.

The inputs

• text - a STRING (or list of strings). Wire in any of the pack's text-producing outputs. That's it. One input, one output, both called text.

How to install it

It ships as part of ComfyUI_QWEN3VL_API, so there's no separate install. Get the pack:

cd ComfyUI/custom_nodes
git clone https://github.com/flywhale-666/ComfyUI_QWEN3VL_API
cd ComfyUI_QWEN3VL_API
pip install -r requirements.txt

Restart ComfyUI. The JavaScript extension only registers itself for nodes named QWEN_TextDisplay, so it won't touch anything else in your graph.

Troubleshooting

• The node shows nothing after a run - the display is driven by the JS extension, which needs the pack's web directory in place. If you copied the pack manually, make sure you cloned the whole repo; missing the web folder means the widgets never appear even though the node runs.
• It doesn't display a single-string output the way you expected - the input is list-oriented by design, so a plain string gets wrapped in a one-item list. Works fine, just know that's what's happening.
• The node balloons in size - that's the feature. Long model outputs make the node tall; the code deliberately doesn't shrink below your manually set width. Re-run with a shorter prompt or trim the text before it lands here.

The one genuinely annoying thing: it's an output node, so ComfyUI will treat its presence as "this graph produces a result." That's exactly what you want if it's your terminal node, and slightly surprising if you drop one in the middle of a workflow. Keep it at the end of the text path and it behaves itself.