Float slider

Here's a tiny node that solves a real, specific annoyance: a slider that outputs a float from 0 to 1. That's it. That's the whole node. One widget, one value, one output - and yet it's genuinely useful, because a lot of ComfyUI inputs are floats in exactly that range and a proper slider beats typing numbers into a box.

The most common victims: denoise strength, IP-Adapter weight, ControlNet strength, the weight inputs on the various blend/fade nodes. With this node, you drag one slider and that value fans out to every input you wire it into - which is the whole point of a value node in the plumbing layer: one source, many consumers, no retyping. The README even frames it as a drop-in replacement for the slider from 11dogzi/Comfyui-ergouzi-Nodes, so if you recognize it from there, this is the same idea in a smaller pack.

The one input that matters

weight, a FLOAT from 0 to 1 with a step of 0.01, default 1, displayed as a slider. Output: a FLOAT with that value. The node description in the source is literally "simple slider from 0 to 1."

A couple of practical notes. It only does 0–1 - if you need a wider range (say a CFG slider up to 10), this isn't the node; you'd reach for a generic Primitive float or one of the range sliders from a bigger pack. And because it's just a value node, nothing downstream re-runs unless you move the slider - which is what you want. The step of 0.01 is a good default; fine enough for denoise work, coarse enough to actually drag.

Install

ComfyUI Manager: search ComfyUI_Gayrat and install, then restart. By hand:

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

It lives under the Gayrat menu. No dependencies, no models, no network - pure UI convenience. It comes bundled in a personal grab-bag pack that pulls in a dozen other nodes you may not care about; if you only want the slider, you can copy float_slider.py into a custom node folder of your own, or just live with the pack. For one honest take: this is the kind of node that sounds too trivial to install a pack for, and then you find yourself rebuilding it by hand in every workflow anyway. Now you don't have to.