Ksampler Scheduler Arg (watdafox)

Every KSampler in your graph already has a scheduler dropdown built in. So why would you want a second node whose entire job is picking a scheduler?

Two reasons, and they're the reasons most "parameter pack" nodes exist in the first place. First, one source of truth. If you run a base pass, an upscale pass and a detailer pass, and you want all three to use karras, you can wire one Ksampler Scheduler Arg into all of them instead of remembering to change three dropdowns every time you experiment. Second, and sneakier: KSampler only exposes its scheduler as a dropdown that lives inside the node - you can't read it, log it, or route it. This node's second output, str_scheduler, is a plain STRING you can pipe into a Show Text / text-saving node so your workflow's metadata records which scheduler actually ran. That's the quiet killer feature here.

How it works

It's a pure passthrough. Open the source and the whole thing is a few lines: a required scheduler combo built from ComfyUI's own comfy.samplers.KSampler.SCHEDULERS list (so it always matches what your KSampler understands - simple, sgm_uniform, karras, exponential, ddim_uniform, beta, normal, linear_quadratic, kl_optimal), and RETURN_TYPES = (scheduler_list, "STRING"). The execute function just returns (scheduler, scheduler). That's it. No API calls, no state, no randomness - the name is a lie in the best way: it's boring on purpose.

The two inputs/outputs that matter

• scheduler - the combo. Pick the one you want.
• Outputs scheduler (feeds the scheduler input of any KSampler / KSampler Advanced) and str_scheduler (the same value as a STRING for logging or routing into text nodes).

That's the whole interface. It appears under the watdafox/parameter menu, and it belongs to the same family as the pack's Ksampler Sampler Arg and Detailer Scheduler Arg - grab the three of them and you've got every sampling knob as a reusable, wire-able value.

Install

cd ComfyUI/custom_nodes
git clone https://github.com/OhSeongHyeon/comfyui-watdafox-nodes.git

Restart ComfyUI. Or install it through ComfyUI Manager by searching "comfyui-watdafox-nodes". No extra pip packages and no model downloads - the pack's dependency list is empty and everything it imports ships with ComfyUI.

When it bites

Honestly, the main trap is convincing yourself you need it. If you have a single KSampler and you don't care about logging, the stock dropdown is fine and this node is ceremony. But the moment you start building multi-pass workflows and repeating your settings across samplers, or you want run metadata that actually says what scheduler ran, this is the kind of node that quietly prevents an afternoon of "which graph had the good settings."