Float | Basic

Some nodes just exist to hand you a number you control, and FloatInput ("Float | Basic") is exactly that. It's a single widget - a decimal value with a slider - that outputs a FLOAT you can wire into anything that takes a float. Sounds almost too simple to write about, but in practice it's the difference between hard-coding a 0.65 denoise into one specific node and having one knob that drives five nodes from the same place.

You get one input, value: a FLOAT with a default of 0, a range of roughly ±1e12, and a step of 0.001. That last bit matters more than it looks. The step size is what your mouse does when you drag the widget, and 0.001 is fine-grained enough for things like denoise, CFG, or LoRA strength without the slider feeling twitchy. The single output is FLOAT.

Why you'd reach for it

ComfyUI is happy to inline a float into a node's widget, and for a one-off value that's usually the right call. You reach for a standalone input node when the same number needs to appear in several places, or when you want it exposed as an obvious, findable knob rather than buried in a node's properties. It also plays nicely with this pack's union types: a FLOAT wire will connect into any INT-or-FLOAT socket, so you can feed this straight into BasicMath, NumberLerp, or a sampler without a conversion in between.

It's not the most exciting node in the pack - it's the most boring one, and that's the point. Primitives like this are the load-bearing parts of a clean graph.

Installing it

Same story as every node in this pack - the Basic Math pack is pure Python with no dependencies and no model downloads:

cd ComfyUI/custom_nodes
git clone https://github.com/akatz-ai/ComfyUI-Basic-Math

Restart ComfyUI and look under Basic Math ➕ → Primitives. Or use ComfyUI Manager → Install Custom Nodes → search "ComfyUI-Basic-Math".

The gotchas

If you need finer control than 0.001 per drag step, the same pack ships a PreciseFloatInput with a step of 1e-10 - that's the one for micro-tuning. And if the value you're entering keeps rounding to three decimals, that's the step size doing its job, not a bug; you can still type a full-precision value directly into the field.

Watch the menu name too: it's listed as "Float | Basic", with the | Basic suffix this pack appends to every node so it doesn't collide with the many other float inputs in the ecosystem.