Sampler Selector

This is the simplest node in the whole pack, so let's not pretend it's more than it is: Sampler Selector is a dropdown that outputs the sampler name you picked. That's it. One input, one output.

So why does it exist? Because of a ComfyUI typing quirk. A lot of core nodes (and third-party ones) accept a sampler as a loose value - the socket type is *, meaning "anything goes" - and there's no built-in dropdown widget for that kind of socket. This node gives you a proper ComfyUI combo dropdown (populated straight from ComfyUI's own sampler list, so euler, dpmpp_2m, uni_pc, and everything else a stock KSampler offers is right there) and emits the selection as that flexible Any-type value. Wire it into any sampler-name socket and the connection just works.

It's also the building block for the pack's bigger ideas. A Sampler Selector output is a plain string, which means it can ride along inside the pack's metadata system - save an image, and the sampler choice is stored in the A1111 infotext; load it back and the choice is restored. In that world, having sampler selection be its own explicit node (instead of buried inside a KSampler widget) is what lets the metadata round-trip carry the value at all.

Inputs and outputs

• sampler (default euler) - the dropdown. The value is validated against ComfyUI's sampler names before it leaves the node; an invalid pick silently yields nothing rather than a corrupt value.
• sampler (output) - the chosen name as an Any-type value. Feed it into a KSamplerSelect, or anywhere else a sampler name socket appears.

Install

Pack-standard:

cd ComfyUI/custom_nodes
git clone https://github.com/kinorax/comfyui-info-prompt-toolkit.git
cd comfyui-info-prompt-toolkit
pip install -r requirements.txt

Or ComfyUI Manager → search "ComfyUI-Info-Prompt-Toolkit" → install → restart. Nothing else required.

When you'd actually use it

Honestly? If you're hand-wiring one sampler node, you don't need this - the KSampler's own widget is closer to the action. It earns its place in two scenarios: (1) you're building a reusable workflow where the same sampler feeds several places and you want one knob to control all of them, and (2) you're working inside this pack's selector-and-metadata idiom, where choices are meant to be first-class values that persist. For a beginner, think of it as "a sampler picker you can hand to a wire." It won't teach you which sampler to pick - that's on you and your model - but it's a clean, typed way to make the choice portable.