Sampler Select name
One dropdown to rule every sampler in your graph
- sampler
Sampler choices are a pain to manage when you've got several KSamplers in one workflow. Every one of them carries its own sampler_name dropdown, and if you're A/B testing "what does dpmpp_2m look like here" you end up flipping three widgets by hand and hoping you remembered them all. PgSamplerSelectCombo fixes that the dumb way, which is also the reliable way: one dropdown, and the name it produces is wired straight into however many samplers you like.
It's part of PG Nodes' family of "select" nodes, which are all the same pattern - a dropdown of names that outputs the chosen string. This one reads its list from ComfyUI's own KSampler.SAMPLERS registry at import time, so the dropdown isn't a hardcoded copy of something: whatever samplers your ComfyUI actually knows about are the options you get. Install a pack like RES4LYF that registers extra samplers and they show up here automatically - that's why you'll see res_multistep, euler_cfg_pp and the other post-2025 high-order solvers sitting in the list next to the classics.
How it works
Boring on purpose. The node takes sampler_name (a dropdown with 44 entries by default) and returns it as its output - the output type is literally the list of sampler names, which is the trick that lets it snap onto a KSampler's sampler_name input instead of you retyping the string into a widget. The KSampler's sampler input is itself an enum, so ComfyUI accepts the connection and the one choice drives every downstream sampler that shares it.
What it does not do is return a sampler object. That's the sibling node PgSamplerSelect (non-Combo), which tries to hand you a live SAMPLER object via comfy.samplers.sampler_object() for custom sampling calls. The Combo variant is specifically the one for the standard KSampler name slot. Grab the wrong one and the wire won't connect - that's the tell.
Inputs and outputs
sampler_name- the dropdown. Defaulteuler; the full list reflects whateverKSampler.SAMPLERScontains in your install at load time.- Output
sampler- the chosen name as a string, typed to the sampler-name enum so it plugs into KSampler.
One knob to mention, because it bites people: the list is materialized at import time. If you install a new sampler pack, restart ComfyUI before you expect the new names in the dropdown.
Install
Standard PG Nodes install - ComfyUI Manager, search "PG Nodes", install, restart. Or:
cd ComfyUI/custom_nodes
git clone https://github.com/GizmoR13/PG-Nodes
No models, no dependencies beyond ComfyUI core.
When to reach for it
Honestly? For one sampler in a small graph, just use the KSampler's own dropdown - this node is ceremony. It earns its place when you have parallel samplers (a detail pass and a base pass, or a tiled upscale loop) that must stay in lockstep, or when you want the sampler choice sitting in one visible place so the "what did we run with" question has a single answer. Wire the output into every KSampler's sampler_name and you'll never drift again.
Inputs (1)
| Name | Type | Default | Description |
|---|---|---|---|
| sampler_name | COMBO | 44 options: euler, euler_cfg_pp, euler_ancestral, euler_ancestral_cfg_pp, heun, heunpp2, +38 |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| sampler | euler,euler_cfg_pp,euler_ancestral,euler_ancestral_cfg_pp,heun,heunpp2,exp_heun_2_x0,exp_heun_2_x0_sde,dpm_2,dpm_2_ancestral,lms,dpm_fast,dpm_adaptive,dpmpp_2s_ancestral,dpmpp_2s_ancestral_cfg_pp,dpmpp_sde,dpmpp_sde_gpu,dpmpp_2m,dpmpp_2m_cfg_pp,dpmpp_2m_sde,dpmpp_2m_sde_gpu,dpmpp_2m_sde_heun,dpmpp_2m_sde_heun_gpu,dpmpp_3m_sde,dpmpp_3m_sde_gpu,ddpm,lcm,ipndm,ipndm_v,deis,res_multistep,res_multistep_cfg_pp,res_multistep_ancestral,res_multistep_ancestral_cfg_pp,gradient_estimation,gradient_estimation_cfg_pp,er_sde,seeds_2,seeds_3,sa_solver,sa_solver_pece,ddim,uni_pc,uni_pc_bh2 | — |