📄 LLM Text Display

LLM outputs are long. Really long - full paragraphs, dialogue blocks, structured rewrites - and the default ComfyUI text preview widget is not built for that. 📄 LLM Text Display is the pack's answer: a proper read-only textarea that lives on the node, resizes with it, scrolls, and comes with Copy and Select All buttons. If you're running the 🚀 API runners, this is the node you should be reading their answers in.

It's an output node: one input, no outputs to wire onward. Feed it the runner's response_text and it displays whatever came back. It takes any STRING, so it'll happily show text from anywhere in the graph, but in this pack it's the natural partner for the runner outputs.

What you get on the canvas

The magic here is in the frontend JS, not the Python. The backend is a two-liner that echoes the text up as a UI message. The frontend then draws a DOM textarea widget with the good stuff:

• A resizable panel - drag the node bigger and the text area grows with it, so a 5,000-character response stops being a scroll torture test.
• A monospace, dark-theme textarea with word-wrap, so prompt-engineered output reads like it should.
• Select All and Copy buttons - one click copies the whole response to the clipboard (with a "✅ Copied!" confirmation). This is the killer feature for anyone who's been manually selecting a wall of text from a tiny preview.
• It's read-only, so you can't fat-finger edits into it.

When a new run lands, the node updates the textarea and scrolls it back to the top so you see the beginning, not the tail.

Wiring it in

The single input, text (STRING), is marked forceInput in the source - wire it from a node's output rather than typing into it. The obvious chain:

✨ Gemini Configurator → 🚀 Gemini API Runner → 📄 LLM Text Display

Run the workflow, queue the graph, and the response appears on the node. Pair it with the 🔊 LLM Audio Save & Play node and you've got the pack's full read-and-hear output story.

When you'd use something else

For short one-liners, the stock text preview is fine and takes up less canvas. This node earns its place specifically for long-form outputs - which, let's be honest, is most LLM output worth reading. If you're doing a lot of copy-paste from your workflow, the Copy button alone justifies the extra node on the canvas.

Install

It ships in ComfyUI-Universal-LLM-Suite, so: ComfyUI Manager → search "Universal LLM Suite", or:

cd ComfyUI/custom_nodes/
git clone https://github.com/alice-ai-wonder/ComfyUI-Universal-LLM-Suite

Restart. The pack auto-installs google-genai, openai, and librosa on first boot even though this node needs none of them - that's just how the pack loads. No API key, no model, no cost: this one's the free end of the suite.