After training over 100 LoRAs with more than 24,000 downloads and 1.5M uses, I've decided to start taking commissions.
Whether you want to replicate an artist's style, capture a specific character, or create something around a concept.
🔥 What can I train?
🎨 Style LoRAs: Artist replication, aesthetic styles, illustration styles
👤 Character LoRAs: OCs, existing characters
💡 Concept LoRAs: Environments, poses, compositions, visual concepts
👗 Clothing: Outfits, accessories, hairstyles, makeup
🌐 Other: If you have something in mind, just ask — I'm open to experimenting
All models are trained on Illustrious (my primary base), though I can discuss other bases depending on your needs (keep in mind my experience with other base models is limited).
My specialty is artistic styles — here's a quick look at some of my work:
⚙️ How does the process work?
1. Contact me: You can reach me through CivitAI's built-in chat, Discord (mexes7), or by sending a commission request on Ko-Fi: https://ko-fi.com/mexes via the commissions section (though I prefer to discuss commissions over chat first).
2. We talk about what you want: Tell me what you're looking for — style, character, concept, and any reference images or examples you have in mind.
3. Dataset: I gather the images, clean them of text, watermarks, and other noise, then caption and prep them for training.
4. Training: I train the LoRA and run quality checks to make sure it works properly across different prompts and checkpoints before delivery.
5. Demo: I generate several images and send them to the client, showing how it behaves across different checkpoints, tags, and weights — including different tag combinations if the LoRA has more than one trigger word.
6. Delivery: Once everything looks good, you receive the LoRA (.safetensors) along with a pack of high-quality images (hires.fix, adetailer, etc.) so you can see the full potential of the model.
💰 Pricing
Standard LoRA
I charge $15 for styles, concepts, and characters. More complex or larger-scale projects may vary. (Also applies to OCs with a pre-made dataset.)
OC without dataset:
I charge $25 for creating an OC without a dataset — meaning I handle the entire process of image generation, editing, and LoRA training from scratch.
Also, if you'd like, we can go through a creative process via Discord to create the character from scratch.
An additional $5 to keep the LoRA private, as I publish all the ones I train on my Civitai profile (OCs always remain private).
🏆 My Experience
- I've been making LoRAs for the community for over 2 years. To date, I've published around 110 LoRAs across different platforms, all of which have been widely used and positively rated.
- I have a series of articles explaining training topics and the tools I use — feel free to check them out to verify my knowledge.
- I've participated in several Bounties on the site and have won approximately 15, the majority being style LoRAs along with some character and concept ones.
- I've already fulfilled commissions through Ko-Fi.
- I've completed 5 paid commissions taken directly from the CivitAI Discord.
You can browse my published work here on CivitAI and judge the quality for yourself before committing.
‼️ Clarifications & Limits
Before ordering:
The minimum I work with is 20 images. If the artist or concept has fewer than that, LoRA quality may suffer.
For characters, I can make an exception down to 15 images, but the same caveat applies — quality may drop.
About payment:
Full payment is required upfront — this is to avoid issues.
I accept payments via PayPal and Ko-fi.
About the final result:
A LoRA being 1:1 to what was requested is difficult, if not impossible — AI is inherently variable by nature.
Very small or minute details are hard to replicate precisely due to how AI works (or at least how SDXL: Illustrious/Pony works), so please don't come after me if that tiny star on the character's pants doesn't show up perfectly.
Content limits:
- Nothing that violates CivitAI's ToS
- No scat (or anything similar)
- No gore
- No vore
📋 FAQ
Can I bring my own dataset?
Yes, but I'm not responsible for the final LoRA quality since it largely depends on the dataset.
How long does it take?
Usually between 1 and 4 days depending on my current queue. I'll give you an estimate when we talk.
Can you train NSFW?
Yes.
Can you train Furry?
Yes.
What if I'm not satisfied with the result?
2 revisions are offered per LoRA. Refunds will only be issued in cases where it's clear the LoRA's objective was not met.
Do you keep the LoRA or publish it?
The LoRA is always published unless you pay an additional $5 to keep it private. (OC LoRAs are always private.)
📬 Ready to commission?
Discord: mexes7
Ko-Fi: https://ko-fi.com/mexes
If you have any questions, leave them in the comments or send me a DM