🌸KiraNugget's Generator / Randomizer Anime Workflow🌸
Full Control. Zero Effort. Maximum Results.
Turn ComfyUI into a fully automated anime generation powerhouse — with deep customization, smart randomization, and seamless editing… all in one workflow.
Check out my new Artstyle lora: 🌸KiraNugget's NuggetStyle Artstyle Lora🌸https://civarchive.com/models/2559128/kiranuggets-nuggetstyle-artstyle-lora?modelVersionId=2875910
🔓 Unlock full potential of the Randomizer feature:
⚠️Make sure to update to Masterlist 2.1 for the last Makeup and Skintones update!⚠️
Make sure to download Masterlists Unlock in model versions and follow instructions inside the workflow.
✨ Why You’ll Love It
🎲 No prompt needed — generate instantly or tweak only what you want
🧠 Smart randomizer system for outfits, styles, scenes & more
🎨 Works with ANY checkpoint, LoRA, or style
✏️ Built-in tag edit mode for precise control when you need it
⚡ What It Does
Generate dynamic, high-quality anime images with:
Fully randomized or controlled prompts
Outfit, hairstyle, and visual variety
Seamless switching between automated and manual workflows
From quick generations to fine-tuned edits — it adapts to your workflow.
🎛️ Built for Power Users (But Easy to Use)
Whether you want:
🎯 Plug-and-play generation
🧪 Deep experimentation
🎨 Consistent character outputs
This workflow gives you complete control without complexity.
🚀 Standout Features
🔀 Advanced modular randomization system
🧩 Expandable lists (fully customizable)
💾 Optional prompt saving for reuse
📦 Integrated LoRA Manager support
🖥️ Optional RTX 4K Upscaling
💡 Perfect For
Anime creators
AI art experimenters
Content creators (Shorts, TikTok, etc.)
Anyone tired of typing prompts 😏
To obtain same artstyle as my Previews:
Checkpoint: https://civarchive.com/models/1034922?modelVersionId=2350111 SweetPoison
Lora:
🌸KiraNugget's NuggetStyle Artstyle Lora🌸https://civarchive.com/models/2559128/kiranuggets-nuggetstyle-artstyle-lora?modelVersionId=2875910
Also check out @Maruhashi's page for amazing sliders and concepts Loras; https://civarchive.com/user/Maruhashi
📙 Full worfklow guide:
https://civarchive.com/articles/29183/how-to-use-kiranuggets-generator-randomizer-anime-workflow
❤️ Try It Out
Download, plug in your favorite models & LoRAs, and start generating instantly.
Warning: You might never go back to manual prompting again.
P.S.: NO! Don't look at the backend! You're not suposed to be there!! 😤 Yeah I know i'ts messy...🙄
P.S.2: Thanks to @darksidewalker for the Amazing checkpoints. It motivated me to make a good Illustrious workflow since I was mostly a Qwen-image-edit user before, (still waiting far a Qwen artstyle lora DaSiWa! 😉
Check his other stuff out too! His wan2.2 Workflows ar amazing and most of all he is amazing! ❤️
V3.0
Lot of prompt improivements.
multiple character with regional prompt.
Description
Cleanup front end for ease of use.
Added Image with Metadata image saver (Prompt Only) for easy posting on civit ai.
FAQ
Comments (10)
good workflow, But there's a problem with the MAIN GIRL node; it stops and doesn't continue. Only when I bypass that node does the flow continue.
It should be fixed now, redownload it I noticed that earlier and updated2.2, let me know it it still doesn't work.
Also, make sure to download and install the MASTERLISTS (download in version: Masterlists Unlock), or disable/reselect ALL the settings in the main front end node, this is due to a Masterlist entry that is selected and is curently not in your lists folder.
Nice workflow! Your master list has two identical entries for makeup_list - guessing one is supposed to be color? Either way, super nice.
Thank you for your comment that is really apraciated! :)
Could you tell me what is the duplicate? I will remove it from the list.
Oh the whole list I see, thanks for letting me know, ;)
I got this error message when loading the json from your image, any ideas on how to fix?: TypeError: (v || []).reduce is not a function
My metadata only include the prompt since my workflow is too big, i can't upload images to civitai if I include the workflow in the data.
I get the same error message unfortunately. Loads a portion of the nodes too, but is missing quite a bit of it. This looked like a really useful workflow too :(.
@chingyjee248 you can try the simplified version, its the same thing but it ghenerate 1 image at the time instead of the 3 all at once, just use the same seed for character consistency