Save 4 hours! We auto-setup your workflow! Free!

Drop your workflow.json — we handle every dependency, custom node, and model. Just open the link and run.

Auto-Setup Workflow Json (Free) Now!

ComfyUI Node: LoRA Stacker V3

Class Name

LoraStackerV3_10

Category
loaders
Author
ussoewwin (Account age: 1245days)
Extension
ComfyUI-NunchakuFluxLoraStacker
Latest Updated
2026-10-25
Github Stars
0.02K

How to Install ComfyUI-NunchakuFluxLoraStacker

Install this extension via the ComfyUI Manager by searching for ComfyUI-NunchakuFluxLoraStacker
  • 1. Click the Manager button in the main menu
  • 2. Select Custom Nodes Manager button
  • 3. Enter ComfyUI-NunchakuFluxLoraStacker in the search bar
After installation, click the Restart button to restart ComfyUI. Then, manually refresh your browser to clear the cache and access the updated list of nodes.

Visit ComfyUI Online for ready-to-use ComfyUI environment

  • Free trial available
  • 16GB VRAM to 80GB VRAM GPU machines
  • 400+ preloaded models/nodes
  • Freedom to upload custom models/nodes
  • 200+ ready-to-run workflows
  • 100% private workspace with up to 200GB storage
  • Dedicated Support

Run ComfyUI Online

LoRA Stacker V3 Description

Stacks up to ten unique LoRAs on model and CLIP in order, with master and per-slot controls.

LoRA Stacker V3:

LoraStackerV3_10 applies a sequence of LoRAs to a diffusion model and its CLIP model, letting you combine several visual influences in one workflow. It supports up to ten slots, processes them in slot order, and gives you both a master switch and individual switches for comparing combinations without removing your selections. It skips empty, disabled, and effectively zero-strength slots. When the same LoRA is selected more than once, it applies only the first eligible occurrence, so stacking duplicate selections will not multiply that LoRA’s effect.

LoRA Stacker V3 Input Parameters:

model

Required MODEL input. Connect the diffusion model you want to modify. Each eligible LoRA is applied to the result of the previous one, starting with this model. There is no numeric range, default, or list of options for this connection.

clip

Required CLIP input. Connect the CLIP model that accompanies your diffusion model. Eligible LoRAs are applied to both the diffusion model and CLIP using the same strength, so this connection is part of the stacked result. There is no numeric range, default, or list of options for this connection.

lora_count

Required INT setting that controls how many slots, starting at slot 1, the node examines. The default is 3; the minimum is 1, the maximum is 10, and the step is 1. For example, a value of 3 means selections in slots 4 through 10 are not processed, even if those slots are enabled.

toggle_all

Required BOOLEAN master switch. Options are true and false; the default is true. When false, no LoRAs are applied, regardless of individual slot settings, and the node returns the incoming model and CLIP unchanged. Set it back to true to allow eligible slots to be processed.

enabled_1

Optional BOOLEAN switch for slot 1. Options are true and false; the default is true. Slot 1 can apply its selected LoRA only when this switch and toggle_all are both true and the slot has a nonzero strength.

lora_name_1

Optional LoRA filename selection for slot 1. The supplied interface lists None as its option; the source also builds the selector from LoRA filenames available in ComfyUI. Selecting None leaves this slot unused. No separate default value is specified.

lora_strength_1

Optional FLOAT strength for slot 1. The default is 1.0; the minimum is -100.0, the maximum is 100.0, and the step is 0.01. The same value is passed to the model and CLIP LoRA application. Values whose absolute magnitude is below 0.00001 are skipped; a negative value is permitted, but its visual effect depends on the LoRA and model.

enabled_2

Optional BOOLEAN switch for slot 2. Options are true and false; the default is true. When false, slot 2 is skipped without changing its filename or strength selection. It also requires toggle_all to be true and lora_count to include slot 2.

lora_name_2

Optional LoRA filename selection for slot 2. The supplied interface lists None as its option; the source also builds the selector from available LoRA filenames. Selecting None skips this slot. No separate default value is specified.

lora_strength_2

Optional FLOAT strength for slot 2, applied equally to the model and CLIP. The default is 1.0; the minimum is -100.0, the maximum is 100.0, and the step is 0.01. An absolute value below 0.00001 causes the slot to be skipped.

enabled_3

Optional BOOLEAN switch for slot 3. Options are true and false; the default is true. Turn it off to exclude slot 3 while retaining its settings. The slot is considered only when lora_count is at least 3 and toggle_all is true.

lora_name_3

Optional LoRA filename selection for slot 3. The supplied interface lists None as its option; the source also builds the selector from available LoRA filenames. None means there is no LoRA to apply in this slot. No separate default value is specified.

lora_strength_3

Optional FLOAT strength for slot 3. The default is 1.0; the minimum is -100.0, the maximum is 100.0, and the step is 0.01. It affects both model and CLIP application, while an absolute value below 0.00001 skips the slot.

enabled_4

Optional BOOLEAN switch for slot 4. Options are true and false; the default is true. Even when enabled, slot 4 is ignored unless lora_count is at least 4 and toggle_all is true.

lora_name_4

Optional LoRA filename selection for slot 4. The supplied interface lists None as its option; the source also builds the selector from available LoRA filenames. Choose None to leave slot 4 inactive. No separate default value is specified.

lora_strength_4

Optional FLOAT strength for slot 4. The default is 1.0; the minimum is -100.0, the maximum is 100.0, and the step is 0.01. The node uses this value for both the model and CLIP; an absolute value below 0.00001 prevents application.

enabled_5

Optional BOOLEAN switch for slot 5. Options are true and false; the default is true. It lets you exclude this selection without clearing it. Slot 5 is processed only if lora_count is at least 5 and the master switch is on.

lora_name_5

Optional LoRA filename selection for slot 5. The supplied interface lists None as its option; the source also builds the selector from available LoRA filenames. Selecting None skips the slot. No separate default value is specified.

lora_strength_5

Optional FLOAT strength for slot 5, used equally for the model and CLIP. The default is 1.0; the minimum is -100.0, the maximum is 100.0, and the step is 0.01. The slot is skipped when the absolute strength is below 0.00001.

enabled_6

Optional BOOLEAN switch for slot 6. Options are true and false; the default is true. Slot 6 must also fall within lora_count and have toggle_all enabled before its LoRA can be applied.

lora_name_6

Optional LoRA filename selection for slot 6. The supplied interface lists None as its option; the source also builds the selector from available LoRA filenames. None leaves this slot unused. No separate default value is specified.

lora_strength_6

Optional FLOAT strength for slot 6. The default is 1.0; the minimum is -100.0, the maximum is 100.0, and the step is 0.01. It is passed to both model and CLIP application; an absolute value below 0.00001 skips the slot.

enabled_7

Optional BOOLEAN switch for slot 7. Options are true and false; the default is true. Turning it off excludes this slot while preserving its other settings. Slot 7 is considered only when lora_count is at least 7 and toggle_all is on.

lora_name_7

Optional LoRA filename selection for slot 7. The supplied interface lists None as its option; the source also builds the selector from available LoRA filenames. None causes the node to skip slot 7. No separate default value is specified.

lora_strength_7

Optional FLOAT strength for slot 7. The default is 1.0; the minimum is -100.0, the maximum is 100.0, and the step is 0.01. This strength applies to both model and CLIP, and the slot is skipped if its absolute value is below 0.00001.

enabled_8

Optional BOOLEAN switch for slot 8. Options are true and false; the default is true. For this slot to affect the result, lora_count must be at least 8 and toggle_all must also be true.

lora_name_8

Optional LoRA filename selection for slot 8. The supplied interface lists None as its option; the source also builds the selector from available LoRA filenames. Select None to leave the slot inactive. No separate default value is specified.

lora_strength_8

Optional FLOAT strength for slot 8, shared between model and CLIP application. The default is 1.0; the minimum is -100.0, the maximum is 100.0, and the step is 0.01. An absolute strength below 0.00001 is treated as inactive.

enabled_9

Optional BOOLEAN switch for slot 9. Options are true and false; the default is true. It controls this slot independently, but cannot override a disabled toggle_all or a lora_count below 9.

lora_name_9

Optional LoRA filename selection for slot 9. The supplied interface lists None as its option; the source also builds the selector from available LoRA filenames. None skips this slot. No separate default value is specified.

lora_strength_9

Optional FLOAT strength for slot 9. The default is 1.0; the minimum is -100.0, the maximum is 100.0, and the step is 0.01. It affects both model and CLIP application; an absolute value below 0.00001 skips the slot.

enabled_10

Optional BOOLEAN switch for slot 10. Options are true and false; the default is true. This last slot is considered only when lora_count is 10 and toggle_all is true.

lora_name_10

Optional LoRA filename selection for slot 10. The supplied interface lists None as its option; the source also builds the selector from available LoRA filenames. Selecting None leaves the slot unused. No separate default value is specified.

lora_strength_10

Optional FLOAT strength for slot 10, passed equally to model and CLIP application. The default is 1.0; the minimum is -100.0, the maximum is 100.0, and the step is 0.01. The slot is skipped when the strength’s absolute value is below 0.00001.

LoRA Stacker V3 Output Parameters:

MODEL

The first output is a MODEL. It is the incoming diffusion model after the eligible, distinct LoRAs have been applied in slot order. Connect it to a downstream node that needs the LoRA-modified diffusion model. If no slot qualifies, the node returns the incoming model.

CLIP

The second output is a CLIP. It is the incoming CLIP model after the same eligible LoRAs have been applied in the same order and at the same strengths as the diffusion model. Use it downstream wherever your workflow needs the matching CLIP model. If no slot qualifies, the node returns the incoming CLIP.

LoRA Stacker V3 Usage Tips:

  • Set lora_count to the number of slots you intend to use. With the default value of 3, selections in slots 4 through 10 have no effect.
  • Use toggle_all to compare the entire selected stack with the incoming model and CLIP. Use enabled_1 through enabled_10 to test individual contributions without clearing filenames or strengths.
  • Start with fewer selected LoRAs when evaluating a look, then add slots deliberately. Eligible LoRAs are applied in slot order, so keeping track of that order makes comparisons easier.
  • Do not select the same filename in multiple slots expecting a stronger combined effect. The node applies only its first eligible occurrence; adjust that occurrence’s strength instead.
  • If you want a slot to have no effect, set its filename to None or turn off its individual switch. A strength of 0 also skips it.

LoRA Stacker V3 Common Errors and Solutions:

LoRA file not found or a path-resolution error

  • Explanation: A selected filename could not be resolved when the node tried to load it. The exact error text comes from ComfyUI’s path-resolution function and is not specified in the provided source.
  • Solution: Check that the selected LoRA still exists in a location ComfyUI recognizes for LoRAs, then refresh the available filenames and select it again.

An error while loading the selected LoRA file

  • Explanation: The node uses comfy.utils.load_torch_file with safe loading enabled. An unreadable, damaged, or unsupported selected file may cause loading to fail; the exact error message depends on the underlying loader.
  • Solution: Verify the selected file and try another known-working LoRA. Replace or reacquire the file if it cannot be loaded.

An error while applying a LoRA to MODEL or CLIP

  • Explanation: The node passes the loaded LoRA, incoming model, incoming CLIP, and strength to ComfyUI’s LoRA application function. The provided source does not define a specific message for failures inside that function.
  • Solution: Test the selected LoRA with the intended model and CLIP, and isolate the failing selection by disabling slots one at a time. Check that both incoming connections come from a suitable model-loading workflow.

A selected LoRA has no visible effect

  • Explanation: This is not necessarily an error. The node skips None selections, slots above lora_count, disabled slots, and strengths with an absolute value below 0.00001. It also skips later occurrences of a filename already selected in an earlier eligible slot.
  • Solution: Check lora_count, toggle_all, the slot’s enabled setting, filename, and strength. If the filename appears more than once, adjust its first eligible slot.

All selected LoRAs appear inactive

  • Explanation: When toggle_all is false, every slot is disabled for that execution and the incoming model and CLIP are returned unchanged.
  • Solution: Set toggle_all to true, then check each slot’s individual switch and selection.

Slots 4 through 10 do not apply

  • Explanation: The default lora_count is 3, so the node does not inspect higher-numbered slots until you increase it.
  • Solution: Set lora_count to at least the number of the highest slot you intend to use, up to 10.

LoRA Stacker V3 Related Nodes

Go back to the extension to check out more related nodes.
ComfyUI-NunchakuFluxLoraStacker
RunComfy
Copyright 2025 RunComfy. All Rights Reserved.

RunComfy is the premier ComfyUI platform, offering ComfyUI online environment and services, along with ComfyUI workflows featuring stunning visuals. RunComfy also provides AI Models, enabling artists to harness the latest AI tools to create incredible art.

LoRA Stacker V3