A curated collection of production-grade ComfyUI workflows paired with custom-trained LoRAs across multiple state-of-the-art diffusion architectures (SDXL, FLUX, Klein, and Z-Image Turbo).
Due to file size limits on GitHub, all heavy model weights (.safetensors) are hosted for free on Hugging Face, while execution workflows (.json) and installation guides are stored directly in this repository.
📁 Repository & Directory Layout
To keep your local ComfyUI setup organized, models and workflows are categorized into dedicated subfolders by architecture:
Additional notes- the loras are all different weights and rank, some may output better quality on certain prompts.
For sdxl loras, recommended checkpoints are wai illustrative, anythingxl and meinamix [but it works with almost all sdxl checkpoints, try it out :)]
Samples-
🤗 Hugging Face Model Downloads
Download the .safetensors model weights directly from the official Hugging Face Repository:
Download the .safetensors files for the architectures you wish to use.
Place them into your local ComfyUI setup under ComfyUI/models/loras/ (you can keep them organized in subfolders like loras/sdxl/, loras/flux/, etc.).
2. Load Workflows
Open your local ComfyUI.
Navigate to the workflows/ folder in this repository.
Drag and drop any .json workflow file directly onto your ComfyUI canvas workspace.
Ensure your Load LoRA nodes are pointed to the correct subfolder path for the model you downloaded.
Click Queue Prompt to generate!
🛠️ Recommended Custom Nodes
To ensure all included workflows run smoothly without missing node errors, make sure you have the following custom nodes installed via ComfyUI Manager:
Workflows contained in this repository are open for personal and commercial use under the MIT License.
Individual LoRA model licenses (e.g., FLUX, SDXL non-commercial restrictions) apply based on their underlying base model terms. Check the Hugging Face repository model card for individual commercial licensing details.
