Back with a different kind of creator update~
This month is not really about announcing one big themed plan. Instead, I want to share a few changes to how I handle my projects, commissions, tools, older LoRAs, and requests.
Most of this comes down to making my workflow more realistic. I still want room for my own projects, paid commissions, experimentation, and improving my models without burying myself under too many promises at once.
Nothing dramatic—just some necessary adjustments so I can keep cooking without setting the kitchen on fire.
✨ TL;DR
I will not follow a fixed monthly theme for anime, manga, or manhwa anymore.
My projects may be more random depending on what I feel like working on, but characters from the same series will still be released gradually in smaller batches.
After three more Gokusotsu Kraken uploads, I will pause that series until future volumes reveal more of the characters I still want to train.
Commissions are now officially open through Ko-fi, with limited slots.
You can still commission me with Buzz by converting the listed Ko-fi price into its Buzz equivalent.
Use code
KIBOU2026for 17% off any commission type during the opening period, while the limited slots are available.I am also changing how I prepare manga datasets: less heavy editing, more deliberate tagging.
CivitTracker received new tracking and user-management features.
I may gradually revisit selected older LoRAs using my current dataset workflow.
Free requests are closing so I can focus my available time on personal projects and paid commissions.
🎲 About My Upload Style Going Forward
I have been thinking about how I want to handle my uploads from now on.
I do not think I can follow a strict monthly theme like “this month is anime only” or “next month is manga only.” My brain just does not work well with that kind of structure. I may suddenly feel more motivated to work on a manga character, a manhwa character, an anime character, or something completely different, and I do not want to force myself to wait for the “correct” month.
So future uploads may be a bit more random.
That said, I still want to keep the same general posting style I have now: working on characters from one series in chunks, then gradually uploading them over time instead of dropping the entire set at once.
It is partly for my sanity, because preparing and posting everything from one project all at once can be exhausting. But honestly, I also just enjoy letting a series breathe a little. Releasing the characters gradually gives me something to look forward to, lets me enjoy the project longer, and makes the whole process feel more satisfying for me.
So the source may change randomly depending on what I feel like working on, but when I start a series, you will probably still see its characters appear in small batches instead of one giant upload dump.
🐙 A Small Pause for Gokusotsu Kraken
After three more Gokusotsu Kraken character uploads, I will pause the series for a while and move on to other projects.
There are still more characters from it that I want to train. The problem is that some of them have not been properly revealed yet, so I do not have enough clear material to prepare them the way I want.
Rather than forcing a LoRA from incomplete references, I would rather wait for new volumes to come out and give those characters a better chance later.
So this is not me dropping the series. It is just waiting for the manga to give me more food first~
☕ Commissions Are Officially Open
After someone contacted me through DM and commissioned a LoRA from me, I finally decided to set up a proper commission page instead of handling everything in a completely improvised way.
I now have a Ko-fi page for commissions.
I also made a commission portfolio where I highlight some of the work I personally feel is worth showing. It also includes the commission details and links to my other platforms, including my Ko-fi commission page and Pixiv.
☕ Open my Ko-fi commission page
You can still commission me using Buzz too. I will use the listed Ko-fi price as the reference, then convert it into the equivalent Buzz amount before the commission starts.
That said, commissions will only have a small number of slots available at a time.
I still want to spend most of my time working on my own LoRA projects, so I do not want to turn commissions into an unlimited queue that prevents me from making anything personal.
🏷️ Opening Discount for This Month
To celebrate the official opening of my commissions, I made a special discount code for this month.
Use code KIBOU2026 when placing a commission through Ko-fi to get 17% off any commission type.
The code is only available during this opening period and will stop working once the limited slots are filled, so this is probably the best time to abuse it a little~
Once the opening slots are taken, commissions will close again until I have enough room to accept more work.
📖 Manga Dataset Prep Update
After getting feedback on my dataset-prep article, I have been changing how I prepare manga datasets.
I am now trying to do less heavy editing and rely more on proper tagging instead. For example, rather than removing every emphasis line or motion line, I can keep them when they do not block important details and tag them properly.
So far, this makes manga prep much easier and lets me keep more usable images without spending forever cleaning every panel.
I will explain this approach properly later in a separate article focused on manga dataset preparation.
📊 CivitTracker Has Also Been Updated
I also updated my personal Civit tracking app, CivitTracker.
It can now track more things in one place, including:
Buzz progress
article growth
generated images from other users
public image galleries in grid view
I also added a user-management menu, so I can manually block, unblock, or whitelist specific users when needed.
The image gallery only tracks public images, of course.
Honestly, this app gives me more dopamine than it probably should. Seeing my model download counts go up always makes me smile a little.
I am also really happy with the generated-image tracker. CivitAI does not notify me when someone posts a new image using one of my LoRAs, so I often do not realize people have made something with them until much later. Now I can check those public generations more easily instead of accidentally missing them.
It is still mainly a personal tool, but it helps me monitor my uploads, articles, and community activity without constantly jumping between different CivitAI pages.
Small app, but very useful for feeding my curiosity and checking how things are going without manually looking through everything one by one.
🔁 Updating Older LoRAs
I am also considering updating some of my older LoRAs over time, especially models made before I had my current dataset-preparation workflow.
For the cutoff point, I will use Anzu (ans) from Kuma Kuma Kuma Bear as the reference.
I will only consider update requests for LoRAs that were released before that model. Anything from that point onward is already closer to my newer workflow, so I would rather focus my time on models that are more likely to benefit from a proper rebuild.
My older workflow was much more experimental. I relied more on premade datasets, auto-tagging, and hoping the training result would survive whatever I accidentally threw into it.
Now I spend much more time on image selection, cleanup, tag consistency, outfit separation, and making the dataset easier to control through prompting.
That does not mean every eligible old LoRA will automatically receive a remake. Some may not have enough source material, some may already be good enough, and some may simply not be worth rebuilding right now.
But if there is a LoRA released before the Anzu model that you would like me to revisit, leave a comment below. It will help me see which older characters people are still interested in.
🚫 Free Requests Are Closing
I need to close the free-request list.
I am sorry to everyone who already placed a request there, but I have realized that I cannot keep treating every request as something I am expected to train for free.
A request may not be picked up for different reasons. Sometimes I am simply not interested enough in the character to spend hours preparing the dataset and training them. Sometimes a good version already exists on the platform, so I do not feel like making another one just for the sake of it. Some requests are also clearly meant for commercial use, which is not something I want to take on for free.
There are also requests I cannot safely accept at all, such as characters or concepts that could put my account at risk because of CivitAI rules or other platform restrictions.
And honestly, repeated requests for the same kind of thing can become overwhelming too. I appreciate that people want specific characters, but I still need to choose projects that I actually have the time, motivation, and room to handle properly.
Training a character is not just pressing the train button. I still need to gather sources, select usable images, clean them, tag them, test the result, make showcase images, and prepare the post. That is a lot of work to commit to when the project is not one I personally want to make.
From now on, I want to keep my available time focused on:
my own LoRA projects
paid commissions
Thank you to everyone who submitted ideas. Even when I cannot take them on, I still appreciate that you thought my work was worth asking for.
✨ Closing
That is all for this update.
Commissions are now officially open, the free-request list is closing, and I have a few workflow changes that should make it easier for me to keep improving what I make behind the scenes.
Feel free to check the commission page if you are interested, leave a comment for an older LoRA you would like to see updated, or just tell me what you think about the changes.
And as always, thank you for supporting my work in any form—comments, collections, generations, commissions, or simply using the models. I really appreciate it.