Useful nodes for ComfyUI, including a hybrid mask generator for simultaneous outpainting/inpainting
JLC ComfyUI Nodes
Repository
https://github.com/Damkohler/jlc-comfyui-nodes
A collection of workflow-focused ComfyUI nodes designed to simplify advanced image generation pipelines. Includes tools for flexible image padding and mask merging to enable inpainting and outpainting in a single pass, structured ControlNet application, sequential LoRA stacking (up to 10 LoRAs), a two LoRA loader with block-weight control, and reusable components for Flux-based workflows and complex image generation pipelines.
---
## Installation
### Using git:
Clone this repository into your **ComfyUI `custom_nodes` directory**.
```
ComfyUI/
└── custom_nodes/
```
Then run:
```
git clone https://github.com/Damkohler/jlc-comfyui-nodes.git
```
Restart **ComfyUI** after installation.
The nodes will appear in the standard ComfyUI node menu under their
appropriate workflow categories.
### Install via ComfyUI Manager
1. Open **ComfyUI**
2. Open **Manager**
3. Search for **JLC ComfyUI Nodes**
4. Click **Install**
---
# Nodes Included
| Node | Purpose |
|-----|--------|
| **JLC Padded Image** | Canvas preparation for inpainting and outpainting workflows |
| **JLC Padded Latent** | Combined padded-image + latent + mask conditioning pipeline |
| **JLC ControlNet Apply** | Simplified ControlNet application node |
| **JLC 10 LoRA Loader Stack** | Sequential loader for up to 10 LoRAs |
| **JLC LoRA Loader (Block Weight)** | Multi-slot LoRA loader with block weight control |
---## Example Workflows
PNG workflows contain the embedded ComfyUI graph and can be dragged directly into the ComfyUI canvas.
# Node Descriptions
## JLC Padded Image
A utility node that prepares images for **inpainting or outpainting**
by placing them on a new canvas with a specified aspect ratio and size.
### Features
- Canvas resizing with aspect ratio control
- Image placement using offset controls
- High-quality **Lanczos resampling**
- Automatic outpaint mask generation
- Optional manual mask merging
- Deterministic padding behavior
Designed to work particularly well with inpainting models such as:
```
flux1-fill-dev
```
---
## JLC Padded Latent
A higher-level workflow node that combines:
- padded image preparation
- outpaint mask generation and merge with inpaint masks
- inpaint conditioning
### Outputs
- conditioned positive prompt
- conditioned negative prompt
- latent image
- mask
- image dimensions
This node simplifies building **reusable inpainting pipelines**.
---
## JLC ControlNet Apply
A streamlined node for applying **ControlNet conditioning**
within a generation pipeline.
### Design Goals
- simplified parameter handling
- improved workflow clarity
- compatibility with Flux-based pipelines
This node adapts the built-in **ComfyUI ControlNet application logic**
for cleaner integration into custom workflows.
---
## JLC 10 LoRA Loader Stack
Applies up to **ten LoRA models sequentially** to a base model.
### Features
Each slot includes:
- selectable LoRA file
- independent strength control
Slots operate independently and are applied **in order**.
Empty slots or strengths of zero are automatically skipped.
### Inspiration
Concept inspired by the **LoRA Loader Stack** design by:
https://github.com/rgthree
---
## JLC LoRA Loader (Block Weight)
A LoRA loader with **block weight support**, allowing detailed control
over how LoRA influence is distributed across model layers.
### Features
- multiple LoRA slots
- independent model and CLIP strengths
- per-slot block weight vectors
- sequential LoRA application
This node is adapted from the implementation found in the
**ComfyUI Inspire Pack** project.
Original project:
https://github.com/ltdrdata/ComfyUI-Inspire-Pack
Released under the **MIT License**.
---