TdxhFloatInput

TdxhFloatInput is a floating-point widget (float_value, default 1.0, range ±100000, step 0.01) with a single FLOAT output that echoes the value back. It's a pass-through node in the most literal sense - no math, no clamping, no trickery. ComfyUI's core Float Input node does the same thing.

The reason it's here is the same reason the whole pack exists: the author wanted a consistent set of data sources for graphs built out of TDXH nodes. TdxhIntInput, TdxhFloatInput, and TdxhStringInput are the pack's matched trio, and if you standardize on the pack, they keep every data entry point looking identical. That's a real benefit when you're reading a workflow you built two months ago, and it's the only real benefit this node has over stock.

Where you'd use it: anywhere a FLOAT input needs to be a visible, editable value on the canvas - a denoise strength, a scale factor, a multiplier you're tuning run-to-run. Because the step is 0.01 and the range is huge, it doubles as a fine-grained constant generator. Feed the output into any FLOAT socket; it plays nicely with the strength inputs on the pack's own LoRA, ControlNet, and CLIP-vision nodes if you want a single shared tuning knob feeding several places at once.

Installing it

ComfyUI Manager → search tdxh_node_comfyui → install → restart, or from a terminal:

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

Restart ComfyUI and it's under TDXH → tdxh_data. Small hobbyist pack, GPL-3.0; the README warns nodes may change between versions, and as with all pass-through nodes that warning is mostly theoretical. No models required.

When to skip it

If you aren't already living in a TDXH workflow, the stock Float Input is identical. This node is a consistency choice, not a capability one - worth grabbing when the graph is already full of the pack, pointless otherwise.