Mickmumpitz
patreon
Consistent Character Creator 3.0. IS HERE!
🕑 Added 2025-08-20 09:33:20 +0000 UTCCreating consistent characters is now easier than ever!
Thanks to a new AI model, we can create amazing character datasets from a single image. Unlike previous versions of my Consistent Character Creator, which required merging different models and custom nodes, this version is based on a single model architecture, making it easy to use and install.
The workflow is also modular! You can duplicate groups and adjust prompts to suit your specific needs and preferences.
I've tested it with both realistic images (for generating virtual influencers or product photography) and stylized ones (such as anime and concept art).
However, the Qwen Image Edit model is only a few days old and has significant room for improvement.
Feel free to join the Consistent Character Creator Discord channel to share your findings, problems, and feature requests with me!
INSTALLATION
- Download the .json workflow file
- Drag and drop it into ComfyUI
- Install missing custom nodes if necessary → Update ALL
- Download the models. You can find the links in the workflow and below.
- Hit “R” to refresh or restart ComfyUI. Double check the model loader nodes and select the correct model from the list!
USING THE WORKFLOW
- Import an image of your character. Both full body and close-up images work equally well.
- Optional: Create a simple one-sentence prompt. Describe the image style and clothing, especially elements not fully visible in your input image.
- Configure the workflow. Choose either the FP8 or GGUF + Lightning Lora configuration.
- I recommend using the GGUF + Lightning Lora configuration!
- To select the correct GGUF version for your system, choose the largest version that fits within your available VRAM.
- The workflow is modular: You can duplicate groups and adjust the prompts!
MODELS
📁 ComfyUI\models\unet\gguf
- Qwen-Image-Lightning-4Steps
📁 ComfyUI/models/lora
- qwen_image_edit_fp8_e4m3fn.safetensors (I recommend using the GGUF version instead!)
https://huggingface.co/Comfy-Org/Qwen-Image-Edit_ComfyUI/tree/main/split_files/diffusion_models
📁 ComfyUI/models/diffusion_models
- CLIP
https://huggingface.co/Comfy-Org/Qwen-Image_ComfyUI/tree/main/split_files/text_encoders
📁 ComfyUI\models\clip\qwen
- VAE
https://huggingface.co/Comfy-Org/Qwen-Image_ComfyUI/tree/main/split_files/vae
📁 ComfyUI\models\vae
Common Errors
- Expression Editor not working - Sometimes the value for crop_factor gets lost during installation. Just manually put in 1.5 as crop_factor where it probably says ”nan“ for you and it should work.
Comments
Jeremiah Everett
What do I do for this? Failed to find the following ComfyRegistry list. The cache may be outdated, or the nodes may have been removed from ComfyRegistry. comfy-core comfy-core comfy-core comfy-core comfy-core comfy-core comfy-core
Åke
I get this error message : Runtime Error 647 (on queue process, top left corner) KSampler shape '[1, 16, 728, 2, 360, 2]' is invalid for input of size 50319360 Any idea what's going wrong?
