I was asked, 'Where did you take that photo of that girl?' Midjourney V8.2's beautiful girls are now indistinguishable from humans. 16 AI image generation prompts
Hello! This is Sakista!
The other day, I posted a note about AI image generation using Midjourney V8.2.
This time, I'm talking about how Midjourney V8.2 has become essential if you're going to use AI models for short movies or business projects from now on.
The content I posted on X the other day
(I've been posting more on X lately! Please check it out if you'd like!!)
短いプロンプトで画像生成AIモデルの美学の違いを検証
— さきすた AI artist (@sakisuta_) August 2, 2026
1:Midjourney V8.2
2:NanoBanana Pro
3:GPTImage2
プロンプト:
無地の白いTシャツを着て、ぱっつん前髪のストレートヘアをした18歳の日本人の女の子のスナップショット。背景は学校のグラウンド。 pic.twitter.com/Gz8LIIXcC4
Just like in the post, I will generate images using simple prompts with Midjourney V8.2.

Here are the prompts
An 18-year-old Japanese woman. Straight long hair with blunt bangs. Wearing a white T-shirt, standing on the shore with the ocean spreading out behind her.
As you can see, regardless of prompt skills, anyone can now generate high-quality images.
Midjourney has a reputation for aesthetics, but the point that was difficult to use in previous versions was
in a word, the "overdone feeling".
The quality was high, but it felt fake somewhere.
To eliminate that, it was necessary to use "--style raw" or "--s 0", or work hard to adjust it with cameras, etc.
In other words, the latest Midjourney V8.2 is...
a natural and realistic photographic expression that has the perfection of a shot taken by a professional photographer, while not feeling like the artificiality peculiar to AI.
!
That is precisely why it is suitable for AI movies and AI models that will likely dominate the market more and more from now on!
▼Please take a look at the delicate expressions unique to Midjourney



Suitable for generating beautiful girls
As I tested a large amount after Midjourney V8.2 was released, what became clear is that it is suitable for generating beautiful girls with a sense of transparency.
Midjourney 8.2 飲料CMで見るような、透明感ある美少女の画像生成が非常に得意になってる
— さきすた AI artist (@sakisuta_) July 31, 2026
GPTImage2、NanoBananaとはそもそも方向性が違うが、画像生成AIの中で一線を画す美的センス
V8から生成スピードも向上していて、画像4枚の生成がわずか2〜3秒。
MJ8.2(画像生成)→GPTImage2 or Nano… pic.twitter.com/RdA8sPblb3
To generate images of the same level with AI image generation models like GPTImage2 or NanoBanana, it is necessary to design the prompts quite meticulously.
With Midjourney V8.2, you don't need complicated prompts; you can generate them in one go.
However, you still want your images to have personality, don't you?
Personality can mean many things, from character traits to color palettes.
So, here are a few ways to bring out that personality.
How to bring out personality in [Midjourney V8.2]
I will explain by genre. For beautiful girls, you can specify things like teenager or young woman. In this note, I will fix the age at 18 as an example.
18-year-old {person}
Before that...
Midjourney is an English-based AI model, and since its training data is based on overseas sources, it has a tendency to produce 'freckles' easily.
If that bothers you, the following negative prompt is recommended.
{prompt} --no freckles
Prompts to maintain the atmosphere of a beautiful girl
Round face
I will use this as a base for generation.
Bringing out personality with hairstyles
Bob with blunt bangs

Prompt
18-year-old Japanese woman. Round face. Blunt bangs covering eyebrows. Black hair bob cut at the jawline. Standing in a narrow alley with white walls. Afternoon light reflecting off the white walls. No shadows on the face. Waist-up shot. --no freckles
Low pigtails

Prompt
18-year-old Japanese woman. Round face. Blunt bangs covering eyebrows. Black hair tied in twin tails below the ears. Thin stray hairs falling around the face. Oversized unbleached shirt. Standing in a bright room with white curtains drawn. Diffused light through the curtains leaves no shadows on the face. Waist-up shot. --no freckles
Straight hair below the shoulders

Prompt
18-year-old Japanese woman. Round face. Blunt bangs covering eyebrows. Straight black hair reaching below the shoulders. White crew-neck T-shirt. Standing on a beach on a cloudy day. Uniform light leaves no shadows on the face. Chest-up shot. --no freckles
Setting the camera
Fixing the model with the aforementioned 'straight hair below the shoulders,' let's set the camera.
Specifying the camera type changes even the 'atmosphere' of the photograph.
Focal length, aperture, and film type are each the deciding factors for texture.
iPhone (everyday snapshot feel)

Prompt
18-year-old Japanese woman. Round face. Blunt bangs covering eyebrows. Straight black hair reaching below the shoulders. White crew-neck T-shirt. Standing on a street in a summer residential area. Utility poles and block walls continue in the background. Uniform light on a slightly cloudy day leaves no shadows on the face. Chest-up shot. Shot with an iPhone ultra-wide 13mm f/2.4 lens. Deep focus. Background buildings and power lines are sharply captured. Slight camera shake. Raw, unprocessed colors. --no freckles
Compact film (the raw feel of the 90s)

Prompt
18-year-old Japanese woman. Round face. Blunt bangs covering eyebrows. Straight black hair reaching below the shoulders. White crew-neck T-shirt. Standing on the grass in a park in the evening. Trees spread out in the background. Soft light before sunset leaves no shadows on the face. Chest-up shot. Photo taken with a disposable film camera. Grainy. Peripheral light falloff. Slight color shift toward green and red. Focus is slightly soft. --no freckles
Fujifilm mirrorless (the texture of a professional capturing everyday life)

Prompt
18-year-old Japanese woman. Round face. Blunt bangs covering eyebrows. Straight black hair reaching below the shoulders. White crew-neck T-shirt. Standing on a street in a summer residential area. Utility poles and block walls continue in the background. Uniform light on a slightly cloudy day leaves no shadows on the face. Chest-up shot. Shot with a Fujifilm X100VI, 23mm f/2. Classic Chrome film simulation. Muted, calm colors. The background is slightly soft, but the outlines are clearly visible. --no freckles
Setting the pose
Fixing the model with the aforementioned 'straight hair below the shoulders' and white T-shirt, let's set the pose.
Specifying a pose changes even the 'story' of the photograph.
The direction of the gaze, the movement of the body, whether still or walking—each determines the meaning of the 'captured moment'.
Composition capturing someone in the middle of walking

Prompt
18-year-old Japanese woman. Round face. Blunt bangs covering the eyebrows. Straight black hair reaching below the shoulders. White crew-neck T-shirt. Walking through a residential street in summer. Gaze is directed slightly ahead in the direction of travel. Utility poles and block walls continue in the background. Even, overcast light with no shadows on the face. Chest-up shot. Shot with an iPhone ultra-wide 13mm f/2.4 lens. Deep focus. Buildings and power lines in the background are sharp. Slight motion blur. Raw, unprocessed colors. --no freckles
A moment sitting on a park swing, not looking at the camera

Prompt
18-year-old Japanese woman. Round face. Blunt bangs covering the eyebrows. Straight black hair reaching below the shoulders. White crew-neck T-shirt. Sitting on a swing in a park in the evening. Face is not turned toward the camera. Gaze is cast toward the ground to her left. Not making eye contact with the camera. Trees spread out in the background. Twilight light just after sunset with no shadows on the face. Chest-up shot. Photo taken with a disposable film camera. Grainy texture. Vignetting. Slight color shift toward green and red. Focus is slightly soft. --no freckles, eye contact, looking at camera
Composition leaning against the outer wall of a street cafe, looking straight at the camera

Prompt
18-year-old Japanese woman. Round face. Blunt bangs covering the eyebrows. Straight black hair reaching below the shoulders. White crew-neck T-shirt. Leaning against the outer wall of a street cafe. Looking straight at the camera. Wooden cafe door and sign in the background. Even, overcast light with no shadows on the face. Chest-up shot. Shot with Fujifilm X100VI, 23mm f/2. Classic Chrome film simulation. Muted, calm colors. The background is slightly soft, but the outlines are clearly visible. --no freckles
Utilizing GPTImage2 / NanoBanana
Using Midjourney, I was able to generate a '0 → 1' model.
However, when generating a model, you have an ideal look in mind, such as
clothing, expression, and hair color.
To achieve that,
from here on, let's utilize GPTImage2 and NanoBanana.
*GPTImage2 is ChatGPT's image generation AI model
*NanoBanana is Google's image generation AI model
Changing fashion


Prompt
Change only the clothes of the woman in @image_1 to @image_2. Keep the hairstyle, face, pose, and background as they are.
Changing fashion and hair color


Prompt
Change only the clothes of the woman in @image_1 to those in @image_2. Change the hair color to a pale lemon yellow. Keep the hairstyle, face, pose, and background as they are.
Change to a smile and look away


Prompt
Change only the clothes of the woman in @image_1 to those in @image_2. Change the expression to a smile. The face is not facing the camera. The gaze is directed away from the camera. Keep the hairstyle, pose, and background as they are.
It's taking shape, but for now, it's only a bust-up composition (upper body).
If it stays like this, consistency cannot be maintained for full-body shots.
Next is how to generate a full-body shot.
Change to a full-body shot
Prepare the material for the bottoms and have her wear them.


Done!
However, the length of the top is a bit subtle...! I will adjust it.
Adjust the length of the top

Done! Looks good!
Prompt
Change the composition of the woman in @image_1 to a full-body shot. Change the bottoms to the skirt in @image_2. Make the hem of the top slightly longer so that it lightly overlaps the waist of the skirt. The skirt should be positioned so that the drawstring part of the waist is hidden and only the scalloped embroidery on the hem is visible. She is wearing white socks and black loafers on her feet. Keep the hairstyle, face, and background as they are.
Next, I will generate a character sheet so that it is easy to use as material.
Generate a character sheet
Creating a character sheet makes it easier to maintain consistency and is useful for creating works.

Prompt
Refer to the woman in @image_1. While maintaining the same hairstyle, same clothing, same location, and same lighting, combine three angles into a single image side-by-side. Place a full-body shot from the front on the left, a full-body shot from behind in the center, and a bust-up shot from the front on the right. All three images should use the same alley background and the same evening light. Place the three images adjacent to each other without gaps.
The material is complete, so finally, let's generate a street snap.
Generate a street snap
Finally, I will use the character sheet to generate a street snap.


Prompt
Refer to the woman in @image_1. Maintain the same hairstyle and clothing. A fashion snapshot set in a residential alleyway at dusk. A shot captured from an unfamiliar angle, deviating from standard compositions.
When creating artwork,
rather than generating a character with a single prompt,
it is often better to generate by separating the work from the base as done here.
I hope this serves as a reference for your creative work!
A short video that a fashion brand could use on social media
Finally,
the video generation AI model Seedance 2.5 has been released, so I generated a 15-second video.
AI動画生成モデルSeedance 2.5を使った15秒生成
— さきすた AI artist (@sakisuta_) August 8, 2026
2.0より明らかに自然な表現💫💫
モデルはMidjourney V8.2
プロンプトは公式+αをスキルにしてから生成しました。理想の結果を求めるならスキル+プロンプト生成がおすすめ。Skillは配布する予定です
最初にゴールも決めると良い
例:
[生成ゴール]… pic.twitter.com/cvumzqnteW
In the next note, I plan to introduce how to generate videos and other useful AI utilization methods for apparel brands!
いいなと思ったら応援しよう!
よろしければサポートお願いします! いただいたサポートはクリエイターとしての活動費に使わせていただきます!