🎨 What does the model do?

This model applies a distinctive semi-realistic painterly art style to AI generations. It blends oil painting techniques with modern digital illustration, producing images that feel like hand-crafted paintings rather than standard AI outputs. It adds visible brushstroke texture, film grain, glossy skin rendering, and cinematic color grading to every generation.

🖼️ What is it for?

Realiqué is a style LoRA designed to transform standard AI generations into painterly semi-realistic illustrations. It is primarily built for generating beautiful female characters in everyday, cozy, and cinematic settings with a warm analog feeling.

✅ What is it good at?

- Female character generation with realistic proportions and detailed skin rendering
- Cozy everyday scenes — kitchens, forests, outdoor settings, bedrooms
- Natural and dappled lighting situations
- Warm cinematic color grading and muted tonal palettes
- Glossy skin with specular highlights and subsurface scattering
- Detailed fabric and clothing rendering
- Atmospheric and moody backgrounds
- Close up and medium shots with strong compositional framing

🎯 What should it be used for?

- Lifestyle and slice-of-life character illustrations
- Cozy and atmospheric scene generation
- Semi-realistic female character portraits
- Cinematic single character compositions
- Fantasy characters in grounded realistic settings
- Fanart of well known characters in a painterly style
- Character concept art with a warm analog aesthetic

❌ What is it bad at?

- Multiple characters in the same scene — quality drops significantly
- Highly stylized anime or cartoon aesthetics — the model pulls toward realism
- Pure abstract or non-figurative generation
- Extreme close up facial detail — skin texture can become muddy
- Dark or purely black backgrounds — the model thrives with environmental context
- Highly complex backgrounds with excessive architectural detail
- Male characters — the model is heavily optimized for female subjects

🚫 What should it NOT be used for?

- Generating multiple characters simultaneously
- Pure anime style generations — use a dedicated anime model instead
- Photorealistic generation without painterly qualities — use a photo model instead
- Complex action scenes with multiple figures
- Architecture or landscape only generations without a character subject
- Generating children or minors in any context

⚖️ Recommended LoRA weight:

0.6 — 0.8 for balanced style application
0.9 — 1.0 for maximum style saturation
0.4 — 0.5 for subtle style blending with other LoRAs