SP_Name_ControlNet

ControlNet is what lets a reference image (an edge map, a depth map, a pose skeleton) decide where things go in your generation while the prompt decides what they are - it's the biggest control unlock the ComfyUI ecosystem has, and most serious workflows use at least one. Loading a ControlNet model is normal territory; what SP_Name_ControlNet does is narrower and specific - pick one from your ControlNet folder and get its filename back as text, nothing loaded.

Why you'd want just the name

ControlNet setups are one of the places people iterate a lot - canny versus depth versus pose, or comparing a few community-trained variants of the same condition type against each other. This node lets you record or branch on which specific control model produced a given output, whether that's tagging filenames in a comparison batch or feeding a switch node that changes downstream behavior based on which ControlNet is in play. It costs nothing to run since it never loads the actual weights, which matters if you're doing this check many times inside a larger conditional graph.

Inputs and outputs that matter

• name - a dropdown built from your ControlNet models folder.
• Output: STRING - the selected filename.

How to install it

Through ComfyUI Manager: search "SP-Nodes," install, restart. Manually:

cd ComfyUI/custom_nodes
git clone https://github.com/bananasss00/ComfyUI-SP-Nodes

Restart ComfyUI once it's cloned. No extra dependency for this node.

Common issues & troubleshooting

Dropdown is empty despite having ControlNet models installed. Check they landed in the folder ComfyUI actually expects for ControlNet weights, and restart if they were added after ComfyUI was already running - this node reads the same folder the real ControlNet Loader does, so if one can see them and the other can't, it's almost always a stale scan rather than a wrong folder.

Newer union or multi-condition ControlNet checkpoints show up oddly named. Union checkpoints (one file covering canny, depth, and pose together) are increasingly common, and their filenames don't always make the covered condition types obvious. That's a naming/packaging fact about the model file itself, not something this node can clarify - check the model's own model-card or README if the filename alone doesn't tell you what condition it handles.

Confusing this with an actual ControlNet Loader + Apply chain. This node stops at the filename. If your workflow needs the ControlNet actually applied to conditioning, you still need the real Load ControlNet Model and Apply ControlNet nodes wired into your positive/negative conditioning - this one exists purely for labeling and logic elsewhere in the graph.