ComfyUI OmniConsistency Nodes

ComfyUI OmniConsistency Nodes is a collection of nodes for ComfyUI that allows you to load and use OmniConsistency models.

https://github.com/showlab/OmniConsistency ComfyUI need 40GB VRAM to run.

Installation

Clone the plugin repository into /path/to/ComfyUI/custom_nodes:

git clone https://github.com/lc03lc/Comfyui_OmniConsistency
cd Comfyui_OmniConsistency
pip install -r requirements.txt

Model Download

Required: OmniConsistency main checkpoint
→ https://huggingface.co/showlab/OmniConsistency/resolve/main/OmniConsistency.safetensors

Optional: OmniConsistency built‑in LoRAs
→ https://huggingface.co/showlab/OmniConsistency/tree/main/LoRAs

Note: You can replace lora_path with any FLUX‑1‑based LoRA—whether it comes from Civitai, LiblibAI, or your own training—just point it to the corresponding .safetensors file.

Node Inputs

Port / Widget	Type	Default	Notes
prompt	STRING (multiline)	<Trigger Words>, …	Free‑text description
spatial_image	IMAGE port	—	One upstream image (HWC float32 0‑1, or any ComfyUI image type)
height / width	INT	1024	Must be multiples of 8
guidance_scale	FLOAT	3.5	CFG scale
num_inference_steps	INT	25	DDIM / DPM steps
seed	INT	42	0 = random
base_model_path	STRING	black-forest-labs/FLUX.1-dev	Local folder or HF repo name
omni_model_path	STRING	/path/to/OmniConsistency.safetensors	Main OmniConsistency weights
lora_path (optional)	STRING	empty	Any Flux.1-based extra LoRA; leave blank to disable

Node Output

IMAGE port → list [torch.Tensor (H, W, C) float32 0‑1]
You can wire it directly to Preview Image, Save Image, Upscale, etc.

Built-in LoRAs Results