ACE++ Character Consistency Workflow

The ACE++ Character Consistency workflow is a powerful tool that enables you to generate consistent characters from a single reference image—no training required. Using ACE++, you can maintain a uniform appearance across multiple angles and poses, making it ideal for animations, comics, illustrated stories, and AI-driven content. This workflow seamlessly integrates with Flux Fill to enhance consistency, ensuring that your character retains the original style throughout different renderings.

Why Choose ACE++?

• No Training Required: Generate consistent character images instantly.
• Highly Reliable Character Retention: Works with multiple angles and poses while keeping character features intact.
• Perfect for Creative Applications: Ideal for AI-generated animations, comics, and digital storytelling.
• Customizable Output: Fine-tune style with prompts or Loras, though optimal results align with the original image style.

Originally developed by Sebastian Kamph, ACE++ Character Consistency leverages instruction-based image creation and editing via context-aware content filling. While this workflow is easily accessible, full credit goes to Sebastian Kamph for his innovative approach. Learn more through his tutorial to master ACE++!

For further details, visit: ACE++ Official Page

How to Use ACE++ Character Consistency Workflow?

Steps to Generate Consistent Character Images with ACE++:

1. Upload Your Character Image in Load Image Group.
2. Enter Your Prompts to customize the output.
3. Click Queue Prompt to render the image.

Your generated character images will be saved in the Outputs > ACE folder.

Image Loading and Processing

• Upload a face image, which will be automatically resized to 1024x1024.
• While other sizes are possible, 1024x1024 ensures optimal results.
• This resolution maintains image manageability and ensures proper padding alignment.
• Fully compatible with Flux generation size standards for the best performance.

Image Generation with ACE++

• Write a descriptive prompt and generate your image.
• Flux guidance should be set high, with a default value of 50.
• Use Euler with 20 steps and CFG set to 1.
• Experiment with settings to find the best results.
• The uploaded image dictates the output style.
• Using additional prompts or Loras can modify the style, but may affect image quality.

Saving Images

Your generated images will be saved in ComfyUI > Outputs folder.

ACE++ Models and Setup

When running ACE++ Character Consistency for the first time, necessary models will be downloaded automatically. Please allow 2-3 minutes for the setup on your RunComfy machine.

Required Model Downloads:

1. Portrait Lora Model
   • Download: comfyui_portrait_lora64.safetensors
   • Place in: /models/loras/
2. Flux Fill FP8 Model
   • Download: Flux Fill FP8
   • Place in: /models/diffusion_models/

Conclusion

With ACE++ Character Consistency, you can effortlessly generate high-quality, consistent characters for animations, illustrations, and AI-powered content creation. This workflow eliminates the need for training, streamlining character generation with precision and ease.

Unlock the full potential of ACE++ and bring your characters to life with unmatched consistency!