Sampler Selector

This is the smallest node in the Sage Utils sampler family and it does exactly one thing: it turns a sampler dropdown into a plain string you can wire somewhere. Pick a sampler like dpmpp_2m and out comes the text dpmpp_2m. That's it. No magic, no validation, no model involved.

Why a node that dumb exists

Because in ComfyUI, a combo box value sitting on a node isn't always reachable by other nodes. If you want the sampler name as data - to embed it in metadata, to log it, to feed it into a node that expects a string sampler name - you need something to hand it over as a string output. That's this node. It's the sampler-flavored cousin of the "Selector" family in the pack, and it pairs naturally with the metadata workflow: Sage Utils' whole deal is making sure the metadata you write down matches what you actually ran, and a sampler name you typed into a string field is easier to get wrong than one you picked from the official list.

The inputs and outputs

One required input: sampler_name, a combo box populated with ComfyUI's actual sampler list (default dpmpp_2m), so you can't mistype a sampler that doesn't exist. One output: sampler, a STRING with the selected name. Wire it into a Construct Metadata node, a text joiner, or anything else that wants the name as text.

Honestly? If you never need your sampler name as a string, you don't need this node. It's a utility for people building logging or metadata plumbing, not a workflow essential. Grab it when the metadata node you're using wants the sampler spelled out and you're tired of keeping two dropdowns in sync.

Installing it

It ships in the Sage Utils pack, which installs the usual way - ComfyUI Manager (search Sage Utils) or:

cd ComfyUI/custom_nodes
git clone https://github.com/arcum42/ComfyUI_SageUtils
cd ComfyUI_SageUtils
pip install -r requirements.txt

Restart ComfyUI. The pack's only pip dependency is dynamicprompts, and there are no model downloads to worry about.

Anything to watch for

Not really - it's a passthrough. The only "gotcha" is expecting it to run the sampler. It doesn't; it just names it. If your goal is actually controlling generation, the node you want is the pack's KSampler w/ Sampler Info (which takes a whole SAMPLER_INFO bundle), not this string output. Different job, different node.