💥 Saya from Majo no Tabitabi
— She is the clingy, emotional, and very Elaina-obsessed witch who brings a more chaotic kind of charm to Majo no Tabitabi. She starts off insecure and lonely, but her meeting with Elaina pushes her to grow into someone more capable and determined. Even after improving herself, she still keeps that dramatic, affectionate, slightly hopeless fangirl energy that makes her funny and memorable. A good character if you enjoy cute witches with big feelings and almost zero chill when Elaina is involved.
>Majo no Tabitabi COLLECTION<
⚠️Turn off your R/X/XXX Filter to see hidden JUICY Loras in >HERE<
🖼️ How I Generate My Images
I made a full article that explains how I generate my images — including my settings, prompting tips, and even my potato PC setup 😅
👉 Check it out here if you’re curious or want to try the same method.
🔞 NSFW gallery (18+)
Check my account here to see the spicy showcase that I can't show here, if any.
🎨 Prompt Tips
These are optional starting points for handling a few behaviors of this version. You may still need some retries depending on the checkpoint, prompt, and LoRA strength.
💇 Getting the medium hairstyle
If Saya keeps appearing with short hair when you are trying to generate her medium bob-cut-like hairstyle, try suppressing the short-hair features.
🚫 Negative prompt: short_hair, cowlick
You may also want to avoid prompting side_braid, since it is more strongly associated with her short hairstyle in the training dataset.
🧥 Suppressing the default outfit
Some parts of her default outfit, especially the cloak and collar structure, may remain visible even when you prompt a different outfit.
When changing her clothes, place the unwanted default outfit tags in the negative prompt.
Example:
🚫 Negative prompt: (collared_cloak:1.5), blazer, necktie, cross_tie
Adjust the list based on which pieces keep returning. You do not always need to suppress every part of the outfit at once.
⚙️ LoRA weight and CFG
This version does not seem to have one perfect setting for every prompt, so experimenting with the balance between CFG and LoRA strength is recommended.
Suggested starting ranges:
CFG scale:
4–6LoRA weight:
0.7–1
Sometimes CFG 5 may produce a cleaner result, while another prompt may work better at CFG 6. The same applies to LoRA strength: lowering it can improve outfit flexibility, while raising it may help preserve her face and other character details.
Treat these values as a mix-and-match starting point rather than strict requirements.
⚖️ LoRA strength and outfit flexibility
A weight of 1 can still work with a very different outfit, especially when the new outfit is clearly different from her default one.
Only lower the LoRA weight gradually when outfit details keep leaking even after using clear outfit tags and negative prompts. Lowering it too much may also weaken her face, hair, and overall character identity
📊 Lora Project
If you want to know what characters that I'm covering next you can check these link :
👉 Personal Project
👉 Commision Project
If this model sparked something in you, show some love with a like, review, or buzz.
Now go—
Be wild
Description
🎬 Anime screenshots with official art support
This LoRA was trained mainly using anime screenshots, with some official artwork added for extra support.💇 Two hairstyle variations
This version can depict both of Saya’s hairstyles:
- Her shorter hairstyle
- Her medium-length hairstyle, which has a bob-cut-like shape
The medium hairstyle may be less consistent depending on the checkpoint, prompt, and LoRA strength.🪢 Side braid consistency
The side_braid appears mainly in images where Saya has short hair. Because there are not enough examples combining it with her medium hairstyle, the braid may become inconsistent when both are prompted together.
You may need some retries, or you may get more stable results by not forcing the braid when using her medium hairstyle.🧥 Outfit pieces are individually tagged
The individual pieces of her outfit were tagged separately to provide better prompt control.
However, you should still be mindful when prompting her from angles where certain clothing pieces would normally be hidden. Forcing every outfit tag from every angle may cause details to appear where they should not be visible.👗 Limited outfit separation
I tried to separate her inner dress shirt from the short outer robe or dress worn over it. However, there are no source images showing her wearing the inner layer by itself, so there was not enough visual evidence to teach that variation properly.
Because of that, this version is mainly flexible in whether her witch accessories are included, such as her cape. Removing or changing the entire outer outfit structure may be much less reliable.As usual, results can still depend on the checkpoint, prompt, LoRA strength, CFG scale, and generation settings.