TdxhIntInput

TdxhIntInput is exactly what it says: an integer widget (int_value, range −100000 to 100000, step 1, default 1) that spits the number back out as an INT output. No computation, no conversion, no hidden behavior. It is the same job ComfyUI's core "Int Input" node does, and if you already use that one, you know this one.

So why does it exist? Because the TDXH pack's philosophy is "fewer nodes, consistent UI." The author wrote a matching trio - TdxhIntInput, TdxhFloatInput, TdxhStringInput - so that if you're building a workflow entirely out of this pack, every data source looks and behaves the same. It's a consistency play, not a capability one. If you're already in a graph full of TDXH nodes, grabbing the pack's own input keeps the UI uniform. If you're not, the core node is identical.

Where you'd actually drop it in: feeding a fixed batch count, a step count, a seed-ish constant, or any INT input that you want visible and editable on the canvas rather than buried in a node's widget list. The generous range means it can serve as a simple named constant that you can clamp a value to without reaching for a math node.

Installing it

One install gets the whole pack. ComfyUI Manager → search tdxh_node_comfyui → install → restart, or:

cd ComfyUI/custom_nodes
git clone https://github.com/youyegit/tdxh_node_comfyui

Restart ComfyUI and it lives under TDXH → tdxh_data. It's a small hobbyist pack (GPL-3.0) and the README warns that node behavior can change between versions - for a pass-through node like this, that warning is about as relevant as it gets, which is to say not very. No models, no dependencies beyond the pack's single sentencepiece line.

When to skip it

If you're not standardizing on this pack, ComfyUI's stock Int Input is byte-for-byte the same idea. Reach for TdxhIntInput because the graph is already TDXH-flavored, not because it does something new.