Preview Text (Multiline) 🍕🅩🅕

You've got a wall of prompt text - maybe built by concatenation, maybe assembled from a wildcard node, maybe the output of the translation node in this same pack - and you want to actually read the whole thing before it hits your CLIP Text Encode. Preview Text (Multiline) is built for exactly that: it's the sibling of Preview Text sized for the big CLIP-text-style box instead of a short single-line string.

How it works

No transformation happens here, same as the rest of the Preview Text family - it's a tap you splice inline. The value flowing through gets rendered on the node and passed straight on, so you can insert it into the middle of a working chain purely to look at the text without breaking the connection to whatever comes after it.

The inputs and outputs that matter

• text (required, multiline STRING) - the README frames this as taking "text、primitive (clip text) for input," meaning it's meant to pair with the multiline CLIP-text-style primitive widget rather than a plain short-string primitive, which is what the non-multiline Preview Text is for.
• Output: text - passed through unchanged, so you keep the chain intact.

How to install it

Via ComfyUI Manager: search "ComfyUI_zfkun", install, restart. Manually:

cd ComfyUI/custom_nodes/
git clone https://github.com/zfkun/ComfyUI_zfkun
pip install -r requirements.txt

then restart. Nothing else to set up - no models, no config file, this one just displays text.

Common issues & troubleshooting

You're not sure whether to use this or plain Preview Text. Rule of thumb: if what's flowing through is a short label, a filename, a language code, use Preview Text. If it's an actual prompt - multiple sentences, tag lists, anything you'd normally paste into a CLIP Text Encode box - use this one. They behave identically; the difference is just which primitive widget type they're set up to pair with and how much room the display gives you.

The node is showing an old value. Like any ComfyUI node, it only re-evaluates when its inputs change or you re-run the graph. If you edited something upstream and the preview looks stale, that's a caching quirk of the graph, not a bug in the node - queue the prompt again.

You've got a list of prompt blocks instead of one. This node isn't list-aware - a list connection here will behave the same way it would on plain Preview Text, i.e. not what you want. Use Preview Text (List) if you're previewing a batch rather than a single block.

Leaving these in a "finished" workflow. These preview nodes are genuinely just for you while you're building - there's no harm leaving one wired in after you're done debugging, but if you're sharing the workflow with someone else, it's worth pruning the ones that were only there so you could see intermediate values; they clutter a graph without adding function for someone who just wants to run it.