Nodes/Sage Utils/Quick Lora Stack (x9)
ComfyUI Node

Quick Lora Stack (x9)

Nine LoRAs at once, on/off each, zero clip-weight fiddling

By arcum42·Created 2 years ago·Updated 28 days ago· 33
Quick Lora Stack (x9)
  • lora_stack
  • lora_stack
enabled_1true
lora_1_name
model_1_weight1.00
enabled_2true
lora_2_name
model_2_weight1.00
enabled_3true
lora_3_name
model_3_weight1.00
enabled_4true
lora_4_name
model_4_weight1.00
enabled_5true
lora_5_name
model_5_weight1.00
enabled_6true
lora_6_name
model_6_weight1.00
enabled_7true
lora_7_name
model_7_weight1.00
enabled_8true
lora_8_name
model_8_weight1.00
enabled_9true
lora_9_name
model_9_weight1.00

There's a kind of workflow - heavy Pony character stacks, "kitchen sink" anime prompts, testing five style LoRAs against a base - where six LoRA slots start feeling cramped. Nine is where you stop counting.

Quick Lora Stack (x9) is the maxed-out member of Sage Utils' quick LoRA family: nine LoRAs in one node, each with its own dropdown, weight and on/off switch, all emitting a single lora_stack.

How it works

Same pattern as the x3 and x6 variants, scaled to nine. Each slot gets a triple:

  • enabled_1enabled_9 - boolean per slot, default on.
  • lora_1_namelora_9_name - the LoRA file dropdowns.
  • model_1_weightmodel_9_weight - one weight per LoRA, default 1.0, range -100 to 100.

The optional lora_stack input chains onto an existing stack, so you can grow a stack across nodes in readable chunks. Disabled slots are simply skipped - the stack is built from whatever's enabled, in order, top to bottom.

As with the whole "quick" family, each entry writes its one weight to both the model and clip slots of the stack. The full Lora Stack (x9) sibling exists if you need a separate clip weight per LoRA; the quick version assumes you don't.

Where it fits

Nine slots is the "I'm stacking a full recipe" tier: character, costume, pose, expression, background style, color grading, plus a few experimentals. The booleans earn their keep here more than anywhere - when you're running a comparison grid over which subset of nine LoRAs produces the look you want, flipping switches beats reconnecting wires, and that's precisely the workflow the author built this family for (it grew out of a Reddit request for LoRA loaders with boolean switches).

Feed the output into Load Models + Loras or Lora Stack Loader to apply everything, and into Construct Metadata so the PNG's metadata records the full stack.

Installation

Sage Utils, one install:

cd ComfyUI/custom_nodes
git clone https://github.com/arcum42/ComfyUI_SageUtils
cd ComfyUI_SageUtils
pip install -r requirements.txt

Or ComfyUI Manager → search "Sage Utils" → restart ComfyUI. Only dependency is dynamicprompts.

The honest take

Nine LoRAs is a lot, and sometimes it's too many - weights interact, and debugging which combination broke your composition gets harder the more slots are filled. Start with the booleans all on, then disable one at a time to isolate what's actually doing something. And if you find yourself using five slots out of nine, the x6 node is a tidier fit. But when the full stack is genuinely on the table, this node is the one that holds it without turning your graph into furniture.

CategorySage Utils/lora

Inputs (28)

NameTypeDefaultDescription
enabled_1BOOLEANtrueEnable or disable this LoRA entry in the stack.
lora_1_nameCOMBOSelect a LoRA model to include in this stack entry.
model_1_weightFLOAT1.00-100–100Weight for the LoRA model branch.
enabled_2BOOLEANtrueEnable or disable this LoRA entry in the stack.
lora_2_nameCOMBOSelect a LoRA model to include in this stack entry.
model_2_weightFLOAT1.00-100–100Weight for the LoRA model branch.
enabled_3BOOLEANtrueEnable or disable this LoRA entry in the stack.
lora_3_nameCOMBOSelect a LoRA model to include in this stack entry.
model_3_weightFLOAT1.00-100–100Weight for the LoRA model branch.
enabled_4BOOLEANtrueEnable or disable this LoRA entry in the stack.
lora_4_nameCOMBOSelect a LoRA model to include in this stack entry.
model_4_weightFLOAT1.00-100–100Weight for the LoRA model branch.
enabled_5BOOLEANtrueEnable or disable this LoRA entry in the stack.
lora_5_nameCOMBOSelect a LoRA model to include in this stack entry.
model_5_weightFLOAT1.00-100–100Weight for the LoRA model branch.
enabled_6BOOLEANtrueEnable or disable this LoRA entry in the stack.
lora_6_nameCOMBOSelect a LoRA model to include in this stack entry.
model_6_weightFLOAT1.00-100–100Weight for the LoRA model branch.
enabled_7BOOLEANtrueEnable or disable this LoRA entry in the stack.
lora_7_nameCOMBOSelect a LoRA model to include in this stack entry.
model_7_weightFLOAT1.00-100–100Weight for the LoRA model branch.
enabled_8BOOLEANtrueEnable or disable this LoRA entry in the stack.
lora_8_nameCOMBOSelect a LoRA model to include in this stack entry.
model_8_weightFLOAT1.00-100–100Weight for the LoRA model branch.
enabled_9BOOLEANtrueEnable or disable this LoRA entry in the stack.
lora_9_nameCOMBOSelect a LoRA model to include in this stack entry.
model_9_weightFLOAT1.00-100–100Weight for the LoRA model branch.
lora_stackoptLORA_STACKAn existing LoRA stack to append this entry to.

Outputs (1)

NameTypeDescription
lora_stackLORA_STACKCombined Quick LoRA stack containing the selected entries.