Output Int

ComfyUI_RiceRound isn't a generation pack - it's a layer for turning a working local workflow into a hosted page and callable "cloud node" on riceround.online. RiceRoundOutputIntNode is not one of the nodes that builds that page. It's filed under RiceRound's Output category, and per the pack's own README, that whole family exists to help you troubleshoot a deployment before it goes live - not to sit in the workflow you actually publish. Feed it a whole number, give it a label, and you get back a STRING you can eyeball while you're still building.

Where it sits in the pack

The Input-category sibling here is RiceRoundIntNode - that's the one that becomes an editable integer field on your published page. This node does the opposite job: instead of exposing a value to an end user, it's for you, mid-build, confirming that some integer flowing through your graph is actually what you think it is. Think of it as a print statement wired into ComfyUI's node graph rather than a feature.

The inputs and output that matter

• name (STRING, default "数值" - Chinese for roughly "value" or "number") - a label for whatever you're checking.
• number (INT) - the value being inspected. No default is baked into the schema, so you're setting it explicitly, either by hand or by wiring in something upstream.
• min / max (INT, defaults 0 and 1,000,000) - a much wider band than RiceRoundIntNode's Input-side default of 0–100. That tracks with this being a debug tool rather than something validating end-user input: there's less reason to clip the range tightly when you're just watching a value pass through.
• Output: value (STRING) - not a raw INT. Every node in this pack's Output family normalizes whatever type it receives down to a STRING, which fits a shared debug display more than a data-processing step you'd chain further logic off of.

Installing it

Through ComfyUI Manager: search "RiceRound," install, restart. By hand:

cd ComfyUI/custom_nodes
git clone https://github.com/RiceRound/ComfyUI_RiceRound

then restart ComfyUI. The README doesn't list a requirements.txt or any model download for this pack - it's a packaging and publishing layer sitting on top of a workflow you've already built, so install is quick regardless of which node in it brought you here.

Common issues & troubleshooting

It's still wired into the graph you're about to publish. Take it out. The README is explicit that Input-category nodes are what belong in the workflow you hand to RiceRound Publish, and Output-category nodes - this one included - are for local troubleshooting only. Leaving it in isn't documented as breaking anything, but it's also not the supported path.

You mixed this up with RiceRoundIntNode. Same rough idea (a named integer with bounds), opposite purpose. If your goal is a field the person using your published page can actually change, you want the Input version. If you're just confirming a number mid-build, you're in the right place.

You expected the raw number back and got text. That's correct, not a bug - the STRING output is this whole family's pattern, so build your check accordingly rather than trying to chain the result into something expecting a real INT.