ClownSamplerSelector

This one's small and does exactly what it says: it's a dropdown that outputs a solver name. That's the whole node. You pick a sampler from RES4LYF's full catalog and it hands the chosen name to a Clown sampler's sampler_name input. Think of it as RES4LYF's answer to ComfyUI's KSamplerSelect - a way to expose the solver choice as its own node so you can reroute it, drive it from a primitive, or keep it visible in a big graph instead of buried in a sampler's widgets.

Why bother, when the sampler already has the dropdown

Two honest reasons. One, tidiness: in a complex or reusable workflow you might want the solver choice pulled out where you (or whoever you share it with) can see and change it in one place. Two, and this is the real draw here - this selector exposes the complete list, all 119 entries, organized into categories the inline widget doesn't always surface as cleanly. That's the difference between "the handful of solvers bolded in the readme" and the entire menagerie the pack actually ships.

And it is a menagerie. The list is grouped by solver family:

• multistep/ - res_2m, res_3m, dpmpp_2m, deis variants. Euler-speed, one model call per step. This is where res_2m, the everyday default, lives.
• exponential/ - the big res_2s/3s/5s/10s/16s substep family plus etdrk and lawson variants. Slower, more accurate.
• linear/ - classic Runge-Kutta names: ralston, heun, rk4_4s, dormand-prince, plain euler.
• diag_implicit/ and fully_implicit/ - crouzeix, gauss-legendre, radau, lobatto. These have no stock ComfyUI equivalent at all, and they're expensive.

The input and output

There's exactly one input, sampler_name (default multistep/res_2m), and one output, sampler_name, carrying the selected string. Wire that output into the matching input on ClownSampler_Beta, ClownsharkChainsampler_Beta, or wherever a RES4LYF sampler takes a solver name.

Installing it

ComfyUI Manager: search RES4LYF, install, restart. Or manually:

cd custom_nodes
git clone https://github.com/ClownsharkBatwing/RES4LYF/
cd RES4LYF
pip install -r requirements.txt

pip install inside your venv (or portable Python), restart, hard-refresh F5. No models needed.

Common issues

Don't overthink this node - it selects, it doesn't sample, so on its own it produces nothing visible. It has to feed a sampler. The bigger trap is the paralysis the full list invites: 119 solvers with almost no guidance on which suits what is a real cost of the pack, and the community's honest answer to "which one" is still "it depends on the image." Ignore the temptation to try res_16s. Start at res_2m (fast, works on everything), reach for res_2s or res_3s when you're specifically chasing quality and willing to pay the substep tax, and leave the implicit families for when you have a queue running and time to burn. If the node's missing from your menu, RES4LYF isn't installed.