Quick LoRA Stacker ⚡

The core LoRA loader gives you one LoRA at a time, and chaining three means three loaders wired end to end - workable, but fiddly when all you want is a style LoRA, a character LoRA, and a lighting LoRA stacked at reasonable strengths. Quick LoRA Stacker ⚡ is the pack's fast path: three slots, one model+clip input pair, one combined strength per LoRA, one stack_summary at the end. It's the lighter sibling of the pack's full LoRA Stack, which gives five slots and separate model/clip weights when you need that level of control.

The trade is straightforward: one strength value applied to both model and clip per LoRA, instead of independent weighting. For the common case - stack LoRAs at 0.7, 0.8, 0.5 and go - that's a feature, not a limitation.

How it works

The three lora_N dropdowns come from ComfyUI's LoRA registry (folder_paths.get_filename_list("loras")), the same list the core loader uses, so your files must live where ComfyUI looks - usually ComfyUI/models/loras/. Each slot has its own strength (strength_N, −10 to 10, default 1.0; negative strengths are legal and occasionally useful for suppression). "None" in a slot skips it, so you can use one, two, or three. The stack applies in order and hands back the patched model, clip, and a stack_summary listing what was applied.

The inputs that matter

model and clip in; lora_1/strength_1 through lora_3/strength_3. Wire the output model/clip into your KSampler and CLIP Text Encode exactly as you would from a loader. The stack_summary output is worth glancing at: it confirms which LoRAs actually applied, which beats staring at a silent pipeline wondering why nothing changed.

Where people get burned: stacking three LoRAs at full strength and getting a mushy oversaturated result. Stacking multiplies effects - the community's rule of thumb is that total LoRA influence, not individual strength, is what you're really tuning. Start lower than you think, especially when all three target the same model family.

Installing it

Part of the OmniNodes pack:

cd ComfyUI/custom_nodes/
git clone https://github.com/TensorVizion/OmniNodes

Restart ComfyUI (or search "OmniNodes" in ComfyUI Manager). No extra dependencies - it uses ComfyUI's own comfy.sd LoRA machinery. Registers under TensorVizion/Model Utilities. If a LoRA you expect doesn't appear in the dropdowns, check it's actually in ComfyUI's loras/ folder - the node only sees the registry, not arbitrary paths on disk. If the node itself doesn't show up, restart completely and look for [OmniNodes] ✅ Loaded in the terminal.

For everyday stacking it's the one I'd reach for: three dropdowns, three sliders, done. Reach for the full LoRA Stack when a single strength per LoRA stops being enough.