Scheduler Selector
Pick one scheduler for every sampler in your graph
- scheduler
The scheduler is the noise schedule - it decides how much noise comes off at each step, and it matters more than most people expect. normal is the classic default, karras concentrates denoising in the middle steps (usually better quality at low step counts), and exponential is the other common corner of the map. Every KSampler carries its own scheduler dropdown, so every KSampler carries its own copy. Scheduler Selector is the node that gives you one dropdown for all of them.
What it is
- scheduler - a dropdown populated from ComfyUI's actual scheduler list (the same
normal/karras/exponential/sgm_uniform/simple/betafamily the core nodes use). - Output: scheduler - the selected name as a string, ready to feed a KSampler's
schedulerinput.
That's it. It's the companion to the pack's Sampler Selector: one node for which sampler algorithm, one for which noise schedule. Both exist to keep your "settings" area centralized and your graph uncluttered.
Why you'd reach for it
- Keep multi-sampler graphs in sync. Base pass and refinement pass often share a scheduler. Wire both KSamplers'
schedulerinputs to one selector, change it in one place. - Experiment faster. Toggle between
normalandkarrason one node to feel the difference in a single run, instead of editing dropdowns mid-graph. - Clean settings hubs. If you like a dedicated "settings" region of your workflow (a pattern this pack is built around - it has selectors for models, CLIP types, samplers, and schedulers), this is the piece that feeds your samplers.
The real-world advice from the community still applies: for most models karras at a modest step count is the solid default, and it's the one to reach for when you're testing a new model. The selector just makes switching painless.
Installing it
Part of the Nifty Nodes pack:
cd ComfyUI/custom_nodes
git clone https://github.com/Stibo/comfyui-nifty-nodes
or search "Nifty Nodes" in ComfyUI Manager, then restart. No models or dependencies.
Gotchas
Same caveats as the Sampler Selector: the dropdown only knows what your ComfyUI knows, so update ComfyUI if a scheduler you expect is missing. And remember the output is a plain string - it works with anything that accepts a scheduler name, which is convenient but means there's no type check between you and a typo'd manual override. This pack targets the newer V3 API, so keep ComfyUI updated if the node doesn't appear after install.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| scheduler | COMBO | Select a scheduler to pass as a string to a KSampler node. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| scheduler | * | — |