PROMPT

A multiline text box that outputs a STRING - mechanically identical to the pack's own JN_PrimitiveStringMultiline, and the only real reason to reach for this one specifically is the name. Six months into a big graph, "which text box was my prompt again" is a real question, and a node that's explicitly labeled PROMPT in the menu answers it at a glance instead of making you trace a wire back through a dozen nodes.

It comes from JNComfy (jn-jairo/jn_comfyui), a one-person pack spanning audio, face restoration, and a family of small primitive and logic nodes - this one being about as plain as they get. There's no meaningful community discussion of the pack anywhere, so this is genuinely all there is to know about it beyond the README.

How it works

There's no processing happening inside this node at all - no weighting syntax, no tokenization, nothing. Whatever text you type into value is exactly what comes out on the STRING output. It's a value holder, not a prompt engine; anything you'd normally do to a prompt (attention weighting, LoRA tags, wildcards) still needs to happen in whatever consumes the string downstream - typically a CLIP text encoder.

That's also why it's worth having as a separate node rather than a widget baked into your text encoder directly: a standalone prompt node is a single source of truth you can wire to more than one encoder at once - a positive prompt shared across two samplers in an A/B comparison graph, say - without retyping it or risking the two copies drifting apart.

The inputs and outputs that matter

• value (required, multiline STRING, default empty) - your prompt text.
• Output: STRING - the same text, unchanged.

Installing it

ComfyUI Manager: search "JNComfy". Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/jn-jairo/jn_comfyui

Restart ComfyUI. No models, no dependencies - it's a text box.

Common issues & troubleshooting

Weird literal \n characters or escaped quotes showing up in the text. That's almost always a paste artifact from copying prompt text out of a workflow JSON file rather than something this node introduces - check what's actually in the widget versus what you pasted from.

A very long prompt seems to get ignored past a certain point. That's a limit on whatever CLIP text encoder you're feeding downstream, not this node - most SD/SDXL/Flux text encoders have a token budget per chunk (roughly 75 tokens is the classic SD 1.5/SDXL figure) and will truncate or split silently beyond it. This node itself has no length limit and doesn't warn you either way.

Not sure whether to use this or JN_PrimitiveStringMultiline. They're the same mechanism under different labels. Use this one when the text is genuinely a prompt headed for a text encoder, and the generic multiline node for anything else - a filename template, a JSON snippet, whatever isn't conceptually a prompt.