🐧 Select LoRA Name

Every LoRA loader in ComfyUI carries its own lora_name dropdown, and if you stack several loaders - a base character LoRA, a style LoRA, a detail LoRA at different strengths - that's several places to change the file when you want to swap one of them out. SuperSelectLoraName collapses that to one. It's a single dropdown that outputs the chosen filename, and you feed that one output into the lora_name input of however many loaders you like. Each loader keeps its own strength, but the file itself is now selected in exactly one place.

That's the whole job, and it's a genuinely useful pattern. The workflow-hygiene payoff: swap the LoRA once, and every loader that should share it follows along - no hunting through the graph for the other three dropdowns, no "did I even change that one?" after a session of swapping. If you use rgthree's Power Lora Loader, you already have something similar for the loading side; this node is the lighter, single-file variant that plays nice with plain LoRA loaders and anything else that takes a filename string.

Inputs and output. One input: lora_name, a dropdown populated from your models/loras folder - the exact same list any LoRA loader shows. One output: lora_name, typed as a wildcard (*), which is why it'll connect to basically anything that accepts a string. Feed it to the LoRA loaders' lora_name inputs and leave their strength_model / strength_clip widgets alone. That's the entire node - it just passes the string through.

A note for beginners who run into this and wonder why the output is so eager to connect everywhere: the * type means "any socket," so ComfyUI will happily let you wire it into inputs that are not LoRA names. It'll still work wherever a string belongs, but it's not magic - it can't know that image input over there actually wants a tensor.

Gotchas. The standard ComfyUI one: if you rename or delete a LoRA file, any workflow that selected it will show it as missing, and this node is no exception - the dropdown just won't offer it. And the tooltip-worthy detail is that only the name is shared; strengths stay per-loader by design. That's the feature, not a bug - it's how you run the same LoRA at 0.6 for composition and 1.1 for detail from a single selection.

Install. ComfyUI Manager → search "ComfyUI-SuperNodes", or:

cd ComfyUI/custom_nodes && git clone https://github.com/sonnybox/ComfyUI-SuperNodes

then restart ComfyUI. No models, no dependencies beyond the pack's single matplotlib requirement. Note the pack is written against ComfyUI's newer extension API, so keep ComfyUI updated - on an old build these nodes won't appear at all.