Sign In

jcupi

462

Download

1 variant available

SafeTensor

238.27 MB

Verified:

Type

LoRA

Stats

29

462

267

Reviews

Published

Jan 31, 2026

Base Model

Illustrious

Hash

AutoV2
C828A9FAAA

Trigger Words

jcupi

Created on Civitai

jcupi – Illustrious Character LoRA

Cinematic Character Control • Glasses • Mature Proportions • Identity Stable

Base Model: Illustrious
Trigger:
JCUPI
CLIP Skip: 2


Built With Intent

jcupi was designed for creators who care about control.

Not random aesthetics.
Not slider chaos.
Not “generic anime output.”

She was trained and tested across cinematic, portrait, and high-contrast scenarios to maintain identity stability under pressure — different angles, lighting styles, and composition types.

If you compose with intention, she locks in.


What She Naturally Emphasizes

  • Stable, expressive facial structure (glasses-friendly)

  • Mature, balanced proportions

  • Organic glossy skin (clean specular highlights, no plastic look)

  • Warm golden-hour bias

  • Volumetric depth with floating particles

  • Strong vanishing-point compositions

  • Consistent silhouette across front / ¾ / full body

She leans warm by default.
Increase weight and she amplifies glow and curvature.
Keep her mid-range for identity precision.


Range & Performance

jcupi performs cleanly across:

  • Cinematic hallway shots

  • Portrait lens framing (50–85mm)

  • Soft atmospheric lighting

  • High-contrast close compositions

  • Dynamic body posing

  • Intimate framing without structural collapse

The XY plots above demonstrate smooth strength scaling — identity remains readable from subtle influence to stylized output.

No sudden face drift.
No unpredictable silhouette warping.


Cinematic Soft Lighting

LoRA: 1.05–1.1
CFG: 5–6
Steps: 28–34
Sampler: DPM++ 2M Karras
Smooth gradients, soft highlight rolloff, cleaner depth haze.

Sharper / High-Contrast

LoRA: 0.85–1.0
CFG: 6.5–7
Steps: 40
Sampler: Euler simple
Tighter lines, stronger specular highlights, punchier contrast.


Strength Behavior

Low weight → subtle identity bias
Mid (~1.0) → stable character control
Higher → amplified warmth and curvature

If proportions inflate → lower LoRA slightly
If hair drifts → reinforce hair color in prompt

Detail comes from lighting language, not brute force strength.


Pairs Well With

rim lighting, volumetric rays, shallow depth of field, bloom, dust particles, lens flare, vanishing point, 85mm lens, specular highlights


Creator Notes

This LoRA was built for deliberate composition.

If you throw “masterpiece, best quality” at it, you’ll get something decent.

If you light it intentionally and frame with purpose — it becomes cinematic.

Share your results. I actively refine future releases based on real usage.