🎨 Anima Workflow Pack for ComfyUI

A collection of beginner-friendly Anima workflows for ComfyUI, ranging from simple text-to-image generation to advanced ADetailer workflows with interactive mask editing.

The goal of these workflows is to provide a clean, easy-to-understand setup while still giving advanced users powerful controls when they need them.

📦 Included Workflows

1️⃣ Anima Text-to-Image + 4K PiD Upscaler

A complete text-to-image workflow using Anima Preview 3 with the Anima Turbo LoRA.

Features:

• Optimized for 8–12 steps

• CFG fixed at 1.0 (recommended for Anima Turbo)

• Generates a 1024×1024 image

• Automatically upscales to 4096×4096 using the official PiD Upscaler

• Simple and beginner friendly

2️⃣ Image to 4K PiD Upscaler

Already have an image?

Drop it into this workflow to upscale from 1024×1024 → 4096×4096 using the official PiD Upscaler.

Perfect for:

• Existing Anima renders

• Wallpapers

• Final exports

3️⃣ ADetailer for Anima (Simple)

A lightweight ADetailer workflow.

Pipeline:

Text to Image
→ Detection
→ Detailer
→ Composite

Perfect if you simply want:

• Better faces

• Better hands

• Better eyes

• Better mouth details

No extra setup required.

4️⃣ ADetailer for Anima (Advanced)

This is the workflow I personally use.

It includes everything from the Simple version, plus interactive mask editing for complete control.

Pipeline:

Text to Image
→ Ultralytics Detection
→ Detailer Processor
→ Interactive Mask Editing
→ Detailer KSampler
→ Interactive Composite Mask Editing
→ Composite Back

The two editing stages serve different purposes:

✨ Pre-KSampler Mask

Controls what the AI is allowed to regenerate.

Perfect for:

• Expanding masks

• Removing unwanted areas

• Fixing missed detections

• Pixel-perfect inpainting

✨ Post-KSampler Mask

Controls what gets blended back into the final image.

Perfect for:

• Removing seams

• Hiding artifacts

• Cleaning edges

• Fine-tuning the final composite

This separation gives far more control than traditional ADetailer workflows while avoiding unnecessary re-renders.

📝 Built-in Documentation

Every workflow contains notes explaining:

• Recommended settings

• Denoise

• CFG

• Steps

• Detector settings

• Detailer settings

• Mask editing

• Composite settings

I've also included download notes inside the workflows so you know exactly which models are required.

🧩 Custom Nodes

These workflows only require one custom node pack:

ComfyUI-Ikki-Pack

Repository:

https://github.com/RedsAnalysis/comfyui-ikki-pack

No additional Python libraries are required.

Simply install the node and you're ready to go.

📥 Required Models

Anima

Text Encoders

qwen_3_06b_base

https://huggingface.co/circlestone-labs/Anima/resolve/main/split_files/text_encoders/qwen_3_06b_base.safetensors

gemma_2_2b_it_elm_fp8_scaled

https://huggingface.co/Comfy-Org/PixelDiT/resolve/main/text_encoders/gemma_2_2b_it_elm_fp8_scaled.safetensors

VAE

qwen_image_vae

https://huggingface.co/circlestone-labs/Anima/resolve/main/split_files/vae/qwen_image_vae.safetensors

Diffusion Models

Anima Preview 3

https://huggingface.co/circlestone-labs/Anima/resolve/main/split_files/diffusion_models/anima-preview3-base.safetensors

PiD Upscaler

https://huggingface.co/Comfy-Org/PixelDiT/resolve/main/diffusion_models/pid_1.5_qwenimage_1024_to_4096_4step_int8_convrot.safetensors

Anima Turbo LoRA

https://huggingface.co/circlestone-labs/Anima-Official-LoRAs/resolve/main/anima-turbo-lora-v0.2.safetensors

📂 Model Folder Structure

ComfyUI/
└── models/
    ├── text_encoders/
    │   ├── qwen_3_06b_base.safetensors
    │   └── gemma_2_2b_it_elm_fp8_scaled.safetensors
    │
    ├── vae/
    │   └── qwen_image_vae.safetensors
    │
    ├── diffusion_models/
    │   ├── anima-preview3-base.safetensors
    │   └── pid_1.5_qwenimage_1024_to_4096_4step_int8_convrot.safetensors
    │
    └── loras/
        └── anima-turbo-lora-v0.2.safetensors

🤖 ADetailer Detection Models

Segmentation

Anime Face

https://civarchive.com/models/1076050/adetailer-anime-girl-face-segmentation

Hands

https://civarchive.com/models/329458/hand-detailersegmentation-adetailer

Eyes

https://civarchive.com/models/334668?modelVersionId=465360

Mouth

https://civarchive.com/models/1306938/adetailer-2d-mouth-detection-yolosegmentation

Bounding Box

Face & Hands

https://huggingface.co/Bingsu/adetailer/tree/main

❤️ If You Enjoy These Workflows

If these workflows helped you, consider leaving a like or sharing some of your creations.

I'd genuinely love to see what people make with Anima—whether it's anime wallpapers, illustrations, animations, videos, or anything else.

If you post your work on Instagram, drop your Instagram username in the comments. I'll be happy to check out your art and follow fellow creators in the community.

Happy creating! 🎨✨