Int -> Float

Int -> Float takes an integer and hands you back the equivalent float. It's a two-line node with a single purpose: ComfyUI won't let you plug an INT wire into a FLOAT socket, so the moment you want to run math on a seed, a step count, or any whole number, you need this conversion sitting between them.

How it works

It's float(value) in Python - 7 becomes 7.0. No rounding, no scaling, no tricks: the integer just gains a decimal point, and now it's compatible with every float-typed input and every math node that expects a fractional value. The reason it has to exist is type strictness, not math: the value was already a perfectly good number, but the graph refuses the wire without the explicit conversion.

The classic use is feeding a number that ComfyUI produced as an int - a seed, a frame index, a batch count - into a float math expression or a strength/scale input. Multiply the seed by something, use it as a lerp weight, derive a CFG from a step count. Once it's a float, the whole More Math expression language opens up on it.

Inputs and output

• value - an INT, default 0.
• Output - a FLOAT.

That's all. One input, one output, no options. It's part of the pack's tiny conversion family alongside Float -> Int, Int -> Bool, and Bool -> Int - the pack author clearly didn't want to rebuild these one-liners per project.

Installing it

It ships in More Math (mcDandy/more_math). ComfyUI Manager - search "More Math" - or manually:

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

Restart ComfyUI. Dependencies are antlr4-python3-runtime and torch, no model files. And the pack-wide note applies: it targets a current ComfyUI (newer node API), so if you don't see the nodes after installing, update ComfyUI before troubleshooting anything else.

Any gotchas?

Not really. It's the most boring node in the pack and that's the point. The one thing to remember is direction: this converts to float. If you're going the other way - and you forgot that Float -> Int truncates rather than rounds - that's the node you want, not this one. For conversions in, this is the pair that gets the wires flowing.