Не менять внешность, цвет волос и глаз. Реалистичный портрет 9:16, iPhone 17 Pro, flash photo, lowered exposure, лёгкая зернистость, hyperrealism. Девушка сидит на сером вельветовом диване на коленях, одна рука на спинке, другая касается плеча. Лицо повернуто к плечу, игривый уверенный взгляд в сторону, губы слегка приоткрыты. Очень длинные густые волосы небрежно уложены набок. Glass skin с влажным сиянием, manga lashes, длинные стрелки, glossy dusty-pink lips, тёмно-красный квадратный маникюр. Чёрное мини-платье на тонких бретелях со сборкой, фигура песочные часы. Фон: тёмно-зелёная панельная стена и серый велюровый диван, мягкий приглушённый свет, Y2K Instagram aesthetic.
✦ Use this prompt for free
This AI image prompt was created by PromptSpace and is available to copy and use for free on PromptSpace. At 93 words, this is a detailed prompt that gives AI models precise instructions for generating the exact image you envision.
Copy the prompt
Click the “Copy Prompt” button above to copy the full prompt text to your clipboard.
Open your AI image generator
Open FLUX, Midjourney and paste the prompt into the text field.
Generate your image
Hit generate. For best results, use the default settings first and see how the output looks.
Experiment with variations
Change details like colors, lighting, or composition to make it uniquely yours.
Want your own face in the image?
Upload a clear, well-lit front-facing photo of yourself directly inside your AI tool (look for Image Reference, Face Swap, or IP-Adapter options) — then paste this prompt alongside it to generate images with your own face.
Aspect Ratio
1:1 or 2:3 (Default)
Quality
High / HD
Best For
FLUX, Midjourney
Style
As described in prompt
The high word count gives the model rich context, reducing ambiguity and producing more consistent results. Well-crafted prompts like this one combine subject precision, style vocabulary, and rendering hints — the three pillars of reliable AI image generation.
FLUX is strongly recommended for this prompt. FLUX 1.1 Pro excels at photorealism and handles complex lighting instructions more accurately than older models.
Over-processing is the most common failure — the model may add excessive detail, noise, or unrealistic glow effects. Add 'clean', 'minimal post-processing' or reduce the CFG scale if using Stable Diffusion. Also regenerate 2–3 times before adjusting the prompt.