Control Net Model Input Switch

Exactly what WAS's README says it is: "Switch between two Control Net Model inputs based on a boolean switch." It's the ControlNet member of WAS's larger switch-node family - the same pattern that shows up for Model, CLIP, Image, Latent, VAE, CLIP Vision, Upscale Model, and LoRA inputs elsewhere in the pack. Two inputs of the same type go in, a boolean decides which one comes out, and the rest of your graph downstream doesn't need to care which one actually got selected.

Where this earns a place specifically in a ControlNet workflow: different ControlNet models correspond to different preprocessors - Canny for sharp edges, Depth for spatial layout, OpenPose for figure pose, Tile for upscale detail, and so on - and it's common to want the same downstream "Apply ControlNet" chain to sometimes run one and sometimes another, depending on what the input actually calls for. Rather than maintaining two separate branches of your workflow that only differ in which ControlNet model feeds in, you build one branch and let this node pick between two models based on a condition.

That condition can come from anywhere upstream that produces a boolean - a Text Find checking whether a prompt or filename tag names a specific preprocessor, a Logic AND/OR combination gating on multiple conditions, or a manual toggle if you'd rather flip it by hand between runs. Either way, the workflow only has to exist once; the switch is what makes it conditional instead of duplicated.

Inputs: two CONTROL_NET sockets and one BOOLEAN. Output: a single CONTROL_NET carrying whichever of the two the boolean pointed at, ready to feed into your Apply ControlNet node the same way a single, unswitched ControlNet model would.

Installing it: ComfyUI Manager, search "WAS Node Suite," install, restart - the quick path. Manually: cd ComfyUI/custom_nodes && git clone https://github.com/WASasquatch/was-node-suite-comfyui, then pip install -r requirements.txt (portable: python_embeded\python.exe -s -m pip install -r requirements.txt; manual/venv: activate the venv first), then restart ComfyUI. No model or extra dependency of its own - it's routing logic, nothing more.

Where it bites: the same trap as every switch node in this family - both inputs need something wired into them even if you only ever plan to actually use one in practice, since selecting an unconnected input produces a missing-input error rather than a graceful fallback. Worth double-checking, too, that both ControlNet models you're switching between are actually compatible with the preprocessor feeding your Apply ControlNet step - a model trained for Canny conditioning won't produce sane results if what's arriving is a depth map, and the switch node itself has no way to catch that mismatch for you. Beyond those two things, the standing caveat applies here as everywhere else in the pack: WAS Node Suite has had no active development since December 2023 (the README literally says "Retired"), and the community's recurring complaint is the whole suite failing to import after a ComfyUI update, not individual nodes breaking on their own - check your console for that before troubleshooting this node specifically.