DD Int to Float

ComfyUI is strict about types in a way that infuriates beginners: many nodes demand a FLOAT wire and will refuse an INT one even when the value is exactly the same. DD Int to Float is the two-second fix - an integer in, the same number as a float out. Nothing else. It's the most forgettable node in the pack, and you'll still be glad it's there the fifth time a node rejects your integer.

Why type conversion even matters

In ComfyUI, INT and FLOAT are separate wire types, and a lot of nodes only declare one. Resolutions, steps, and batch sizes are naturally integers; CFG, denoise, and most strengths are floats. The friction appears at the seam: you compute a seed or a step count as an INT, and the node you're feeding wants FLOAT. Without a conversion node you're stuck editing values by hand or reaching for a heavier utility pack. This node is that seam's bridge.

How it works and what it takes

One required input, value (INT, default 0). One output, float_value (FLOAT). The implementation is literally float(value) - no rounding, no clamping, 7 becomes 7.0. It cannot fail and it cannot surprise you, which is the whole appeal.

It pairs cleanly with the rest of the DD math family: feed DD Integer Math into this, then into DD Float Math for a mixed-type pipeline, or keep the result around because some downstream sampler expects FLOAT. And when you need to go the other direction with rounding control, its sibling DD Float to Int covers that.

When to use it (and when not to bother)

Use it when a node throws a type error or refuses to accept an INT input. Don't bother if the destination node already takes INT - converting pointlessly just adds a wire. And a small honest note: this pack is deliberately minimal, so there's no "auto-convert" magic here; you place the node where the type boundary is. That's arguably the better design anyway, because explicit beats hidden whenever you're debugging.

Installing it

It ships in the ComfyUI_DD_Nodes pack by Digital Divas, alongside the other DD math nodes. ComfyUI Manager → search "DD Nodes", or:

cd ComfyUI/custom_nodes
git clone https://github.com/digital-divas-admin/ComfyUI_DD_Nodes.git

Restart ComfyUI. The pack has an empty requirements.txt - zero extra dependencies, zero model downloads. One heads-up when you search Manager: confirm the repo is digital-divas-admin/ComfyUI_DD_Nodes, because an unrelated project also answers to "ComfyUI-DD-Nodes."