Model
Trends
.ai
Model
Trends
.ai
the best open source ai models

Instagirl for Wan 2.2: the realism base layer for video LoRA stacks

Topic: Models
By Captain
Published 2026-10-02
Example renders
Prompts and settings shared by the people who made these renders on Civitai. Pick a render to see how it was made.
Settings
Sampler
res_2s
Steps
12
Guidance
1
Seed
685451517318697
Size
960x1280
Prompt
Instagirl, eye-level medium shot in a messy, sun-drenched hotel room, a young woman with medium-light skin and clear freckles, her shoulder-length straight black hair is tousled from sleep, she is sitting on the edge of the bed with her back to the window, looking over her shoulder at the camera with a sleepy, soft smile, wearing only a man's unbuttoned white dress shirt, her black cord necklace with a small round white pendant rests on her collarbone, intimate charm, kept delicate noise texture, amateur cellphone quality, visible sensor noise, heavy HDR glow, amateur photo, extreme blown-out highlights from the window creating a slight halo effect, deeply crushed shadows on her front.
Negative prompt
色调艳丽,过曝,静态,细节模糊不清,字幕,风格,作品,画作,画面,静止,整体发灰,最差质量,低质量,JPEG压缩残留,丑陋的,残缺的,多余的手指,画得不好的手部,画得不好的脸部,畸形的,毁容的,形态畸形的肢体,手指融合,静止不动的画面,杂乱的背景,三条腿,背景人很多,倒着走, censored, sunburnt skin, rashy skin, red cheeks
Settings
Sampler
res_2s
Steps
12
Guidance
1
Seed
1111197
Size
960x1280
Prompt
Instagirl, amateur photo taken with direct on-camera flash in a dimly lit bathroom, where a teen egirl with long, dark hair in pigtails takes a moody mirror selfie. She holds her phone to obscure one eye, looking at her reflection with a neutral expression, wearing heavy black lipstick and eyeliner. Her skin has a slight oily shine from the harsh flash, which creates deep shadows and visible grain. She is wearing a striped long-sleeve shirt under a faded band t-shirt, a pleated black mini skirt, and a chain belt. On the smudged mirror, written in what looks like pink lipstick, are the words "Instagirl".
Negative prompt
色调艳丽,过曝,静态,细节模糊不清,字幕,风格,作品,画作,画面,静止,整体发灰,最差质量,低质量,JPEG压缩残留,丑陋的,残缺的,多余的手指,画得不好的手部,画得不好的脸部,畸形的,毁容的,形态畸形的肢体,手指融合,静止不动的画面,杂乱的背景,三条腿,背景人很多,倒着走, censored, sunburnt skin, rashy skin, red cheeks
Settings
Steps
8
Size
1152x2048
Prompt
darkelf woman
Negative prompt
色调艳丽,过曝,静态,细节模糊不清,字幕,风格,作品,画作,画面,静止,整体发灰,最差质量,低质量,JPEG压缩残留,丑陋的,残缺的,多余的手指,画得不好的手部,画得不好的脸部,畸形的,毁容的,形态畸形的肢体,手指融合,静止不动的画面,杂乱的背景,三条腿,背景人很多,倒着走.
Settings
Sampler
DPM2
Steps
8
Guidance
5
Size
1152x2048
Prompt
A cinematic instareal, modern lifestyle video featuring a confident, slim 20-year-old woman seated at a rustic wooden table in a cozy, urban coffee shop. She exudes a playful, slightly annoyed yet flirty demeanor, with subtle, expressive glances to the side and occasional looks toward the camera. Large breasts wearing a stylish, form-fitting strapless tube top that accentuates her figure, paired with a delicate, shimmering silver necklace that glints softly under the light. Her actions are natural and engaging: she leans forward slightly to sip her steaming coffee, her fingers lightly tracing the rim of the cup, then casually adjusts her tube top with gentle, deliberate tugs—pulling it slightly downward and repositioning it with a smooth, confident motion. The necklace sways subtly with each adjustment, catching the light and adding a touch of elegance. She occasionally brushes a strand of hair behind her ear, her movements fluid and unhurried, while her playful expression shifts between a coy smirk and a mock-annoyed pout. A sleek laptop sits open on the table, its screen faintly glowing, alongside a ceramic coffee cup with wisps of steam. The coffee shop is bathed in warm, natural daylight streaming through large floor-to-ceiling windows, creating a soft, inviting atmosphere. The background is gently blurred, revealing hints of other patrons, wooden chairs, and cozy decor with exposed brick walls and hanging Edison bulbs. Warm ambient lighting enhances the scene’s charm, casting golden hues across her skin and the table. The video captures her dynamic expressions and movements, culminating in a subtle, coy wink as she gives her tube top one final adjustment—tugging it gently with both hands, her shoulders shifting slightly for emphasis—before looking away with a playful smile. The footage feels intimate and authentic, with slight handheld camera sways for a natural, iPhone-shot aesthetic.
Settings
Size
576x992
Prompt
l3n0v0, Instagirl, photo of a A cute young west european woman with blonde hair and lush green eyes, she has big tits, She is wearing a tanktop with cleavage. shes sitting in a colleage cafeteria, photo from side
Settings
Sampler
res_2s
Steps
12
Guidance
1
Seed
1111179
Size
960x1280
Prompt
Instagirl, a young woman with long, messy blonde hair falling over a pillow, her face looks soft and sleepy, she is peeking one eye open over the edge of a white duvet that is pulled up to her nose, a playful, intimate smile on her lips, wearing a delicate, thin-strapped black silk camisole, extreme close-up from a high angle, as if taken by herself while lying in bed, the scene is filled with the soft, rumpled texture of the white sheets, lit by the gentle, hazy morning light filtering in from a nearby window, soft focus, authentic moment, Instagirl, kept delicate noise texture, dangerous charm, amateur cellphone quality, visible sensor noise, heavy HDR glow, amateur photo, blown-out highlight from the lamp, deeply crushed shadows.
Negative prompt
色调艳丽,过曝,静态,细节模糊不清,字幕,风格,作品,画作,画面,静止,整体发灰,最差质量,低质量,JPEG压缩残留,丑陋的,残缺的,多余的手指,画得不好的手部,画得不好的脸部,畸形的,毁容的,形态畸形的肢体,手指融合,静止不动的画面,杂乱的背景,三条腿,背景人很多,倒着走, censored, sunburnt skin, rashy skin, red cheeks
Settings
Sampler
res_2s
Steps
12
Guidance
1
Seed
1111162
Size
960x1280
Prompt
Instagirl, a young woman with her hair in two high pigtails, dyed half lavender and half mint green, she is winking one eye at the camera and giving a peace sign right next to her cheek, a wide, playful smile on her face, wearing a white collared shirt with a black ribbon tie, tight close-up selfie, taken from a slightly high angle in her bedroom, the background is a collage of anime posters and photos taped to the wall, ring light reflection, high angle selfie, Instagirl, kept delicate noise texture, dangerous charm, amateur cellphone quality, visible sensor noise, heavy HDR glow, amateur photo, blown-out highlight from the lamp, deeply crushed shadows.
Negative prompt
色调艳丽,过曝,静态,细节模糊不清,字幕,风格,作品,画作,画面,静止,整体发灰,最差质量,低质量,JPEG压缩残留,丑陋的,残缺的,多余的手指,画得不好的手部,画得不好的脸部,畸形的,毁容的,形态畸形的肢体,手指融合,静止不动的画面,杂乱的背景,三条腿,背景人很多,倒着走, censored, sunburnt skin, rashy skin, red cheeks
Settings
Steps
12
Guidance
1
Size
720x1280
Prompt
B!MBO!, she has slightly veiny, enormous round fake breasts., A photorealistic full shot of a 23-year-old e-girl with bubblegum pink pigtails and glittery makeup, wearing a tight, pastel purple tube top that clings to her huge, round fake breasts, performing a TikTok dance in a kawaii-themed bedroom. The top’s stretchy fabric emphasizes her curves as she moves, causing a rhythmic, exaggerated bounce of her breasts. The background features pastel fairy lights, anime posters, and a plush-filled bed, creating a soft, playful atmosphere. Warm, diffused lighting with pops of pink and lavender creates gentle shadows that enhance her figure. The fixed camera maintains a centered, dynamic composition, capturing vibrant colors and smooth textures. Aspect ratio 16:9, 22 fps, upbeat pacing with fluid, natural motion.Aspect ratio 16:9, 24 fps, lively pacing with realistic physics.
Instara's Instagirl is a style LoRA for Wan 2.2 T2V designed as the foundation of a LoRA stack: a realistic social-media photo aesthetic on which character LoRAs sit. Learn the trigger, strength, high/low noise variants and prompt ideas.

Overview

Instagirl
WAN 2.2
(v2.5) by Instara is a style
LoRA
for
Wan
Video 2.2 T2V-A14B. It creates the foundational aesthetic of realism (candid, phone-shot, influencer-style footage) upon which you add character or concept LoRAs. The archive includes high-noise and low-noise variants for Wan 2.2's two-stage sampling plus recommended workflows.
Trigger word: Instagirl (some versions use Instacam). Recommended strength 1.0, adjusted in small steps. Danrisi's Lenovo UltraReal for Wan pairs with it.

Reference

Name
Type
What it is
Trigger
Instagirl
Start of prompt.
Strength
1.0
Adjust in small increments.
Variants
high noise / low noise
Load each in its Wan 2.2 stage.
Base
Wan 2.2 T2V 14B
Also used for I2V.
Prompt
candid scene
Describe light, place and an everyday action; adult subjects.
AD
No install required - run AI models on the cloud
BitVector Prism is the easiest way to start: pick a model, type a prompt and generate in a clean web app, with nothing to configure. BitVector is also available on Discord and on the web (SpyGlass).

Step by step

  1. Instagirl, then a candid scene with light described.
  2. Load the high-noise LoRA in the first stage and the low-noise one in the second, both 1.0.
  3. Standard Wan 2.2 steps and resolution (720p).
  4. Add a character LoRA at 0.7 on top.
  5. Add Lenovo UltraReal Wan with l3n0v0 for extra texture.

Examples

Hotel room

Instagirl, eye-level medium shot in a messy sun-drenched hotel room, a young woman in her mid-twenties with medium-light skin and freckles, shoulder-length black hair tousled from sleep, sitting on the edge of the bed stretching, handheld.
Instagirl 1.0 (both stages)

Mirror selfie

Instagirl, amateur video with direct on-camera flash in a dim bathroom, a woman in her twenties with long dark hair takes a moody mirror selfie, phone covering one eye.
Instagirl 1.0

Creative idea: morning routine montage

Instagirl, kitchen at 7 a.m., a woman in her thirties pours coffee, window light, steam, slow handheld push-in.

Tips

  • Direct flash, mirror selfie and sun-drenched room are its native scenes.
  • Describe skin details (freckles) and hair state (tousled).
  • Keep camera moves subtle: handheld, slight drift.
  • Short clips keep faces consistent.

Troubleshooting

Plastic skin

Why it happens
Missing variant in one stage.

How to fix it
Load both high and low noise LoRAs.

Character LoRA overrides the look

Why it happens
Character at 1.0.

How to fix it
Character 0.6-0.7, Instagirl 1.0.

AD
PirateDiffusion
No install required - run AI models on the cloud
PirateDiffusion is Telegram only and built for pros: thousands of models, LoRAs and workflows driven by chat commands, with unlimited generation on a fixed price plan.

Questions

Trigger?

Instagirl (older versions Instacam).

I2V?

Works as the realism layer in I2V too.

Links and sources

Models in this guide

Written by
Captain

Related guides

Models
Lenovo UltraReal for Wan: realistic texture in Wan video
Danrisi's first Wan release of Lenovo UltraReal adds real-camera texture to Wan 14B text-to-video. Learn the l3n0v0 trigger, its limits on facial detail, the Instagirl pairing and ideas.
By Captain
2026-10-01
Models
Aesthetic Quality Modifiers - Masterpiece for Illustrious: stronger quality tags
motimalu's Masterpiece LoRA for Illustrious strengthens masterpiece, best quality and very aesthetic so they produce visibly better composition and finish. Learn the prompt order, weight and ideas.
By Captain
2026-10-01
Models
Height Slider for Krea 2: age-safe height changes
Loraholic's Height Slider changes a subject's height in Krea 2 without altering age. Learn the -4 to 3 range, the note about women at high values and ideas for character line-ups.
By Captain
2026-10-03
Models
Klein Anatomy / Quality Fixer v1.5: a slider that corrects Flux.2 Klein anatomy
l226's Anatomy Fixer for Flux.2 Klein 9B is a concept slider that fixes hands, limbs and glitches at weights 2-3 while keeping the seed. Learn how it works, the right weight and when to use v1.0 instead.
By Captain
2026-09-30

More guides

Models
FLUX.1 Dev: how to prompt it, best settings and creative ideas
A practical guide to FLUX.1 Dev by Black Forest Labs: natural-language prompting, guidance and step settings, LoRA stacking, text rendering and prompt ideas that play to its strengths.
By Captain
2026-09-30
Models
Stable Diffusion XL 1.0: prompts, settings and what it still does best
How to get the most out of the official SDXL 1.0 base model: native resolutions, CFG and sampler settings, the refiner, negative prompts and ideas for styles where SDXL still shines.
By Captain
2026-10-01
Models
Stable Diffusion 1.5: the classic model, prompted properly
Stable Diffusion 1.5 is still worth knowing: the right resolution, CFG and sampler, how to use its huge library of LoRAs and embeddings, and creative prompt ideas suited to a 512-pixel model.
By Captain
2026-10-02
Models
FLUX.2 Dev: prompting the newest Black Forest Labs model
FLUX.2 Dev brings larger prompts, better text, stronger realism and multi-reference editing. This guide covers the turbo workflow, guidance and resolution settings, prompt structure and ideas that use its new strengths.
By Captain
2026-10-03
Models
Qwen-Image: long prompts, perfect text and bilingual posters
Qwen-Image is the model to reach for when the words in the picture matter. Learn how to write its long descriptive prompts, render English and Chinese text accurately, set CFG and steps, and explore layout-heavy creative ideas.
By Captain
2026-09-29
Models
SDXL-Lightning: four-step generation on any SDXL checkpoint
ByteDance's SDXL-Lightning LoRA turns a 30-step SDXL render into a 4-step one. Learn the step counts, the CFG you must use, the sampler, and how to combine it with your favorite checkpoints and style LoRAs.
By Captain
2026-09-30