电影级仙侠提示词海报

Goal: Create a vertical Chinese social-media infographic poster teaching a cinematic Eastern xianxia portrait prompt for {argument name="model name" default="GPT Image-2"}, with a premium dark movie-poster aesthetic and a large rounded white content card.

Canvas: 3:4 vertical poster, dark charcoal-black background with subtle warm bokeh and faint floral haze. Center a large off-white rounded rectangle panel with soft shadow and thin border, leaving a narrow dark margin around it.

Header: At the top, outside or slightly above the white card, place a large pink-white Chinese headline: {argument name="headline text" default="电影级东方仙侠写真提示词"}. Under it, add a smaller white subtitle: {argument name="subtitle text" default="适配 GPT Image-2 完整版"}. Add a tiny centered tagline row beneath: “盛唐宫廷 / 梦幻云海 / 柔焦光影 / 超现实漂浮 / 电影质感”.

Layout: Inside the white panel use a two-column layout. The left column is text-heavy, about 48% width. The right column is a vertical gallery, about 52% width, containing exactly 3 cinematic sample images stacked with rounded corners and thin borders. Each sample has a small pink Chinese label above it: 1) “版本一：云海软榻”, 2) “版本二：仙鹤宫殿”, 3) “版本三：花瓣云端”.

Left column content: Use pink section headings with small circular icons. First section title: “核心提示词（复制使用）”. Under it, set dense black Chinese body text describing a cinematic realistic portrait of an adult ancient Chinese court woman in Tang-style xianxia fashion, including fair skin, delicate features, ornate updo, gold hairpins, pearl tassels, turquoise-blue gauze outer robe, crimson palace bodice, embroidered silk, jewelry, cloud sea, cranes, petals, gemstones, dreamy lighting, bloom, soft glow, 85mm lens, shallow depth of field, cinematic grading, ultra realistic, 8K, HDR, RAW photo, Kodak Portra 400, Unreal cinematic, volumetric light, masterpiece, best quality. Include some English technical phrases mixed into the Chinese text exactly like a prompt sheet.

Negative prompt block: Below the main prompt, add a pink X icon and the heading “负面提示词（避免出现）”. Add a compact paragraph in English: “low quality, blurry, CGI, anime, cartoon, 3D render, doll face, wax skin, modern clothes, text, watermark, bad anatomy, extra fingers, deformed hands, oversaturated, bad proportions, ugly face, jpeg artifacts”.

Keyword block: Add a divider line, then the heading “风格关键词”. Show exactly 10 rounded beige pill tags in two rows: 1) 盛唐宫廷, 2) 东方仙侠, 3) 梦幻云海, 4) 漂浮饰品, 5) 仙鹤, 6) 柔焦胶片, 7) 逆光体积光, 8) 华丽珠宝, 9) 青绿×暗红, 10) 超浅景深.

Usage advice block: Add the heading “使用建议” and exactly 3 bullet points in Chinese: 1) “直接复制【核心提示词】到 GPT Image-2 使用”, 2) “根据需要选择版本一/二/三的描述进行微调”, 3) “可添加 ‘9:16 竖版’ 或 ‘3:4 竖版’ 来控制画面比例”.

Right gallery image details: All 3 images feature the same {argument name="character description" default="young adult Chinese xianxia court beauty with ornate black updo, gold phoenix crown, pearl tassels, fair skin, crimson embroidered bodice, teal translucent silk robe, elaborate jewelry"}. Image 1 shows her reclining among pink clouds inside a dim palace, with floating flowers, warm window light, and surreal dreamy haze. Image 2 shows her seated by a palace window or terrace overlooking a lake with exactly 3 visible cranes, warm sunset mist, dark interior frame, and soft cinematic side light. Image 3 shows her standing or leaning on a balcony above a glowing sea of clouds, petals flying, cranes in the sky, golden backlight, teal and crimson flowing robes, epic romantic atmosphere.

Visual style: Elegant Chinese editorial layout, pale blush-pink typography, clean sans-serif Chinese fonts, cinematic soft contrast, warm highlights, subtle grain, premium Xiaohongshu-style prompt card. Keep the poster information-dense but readable.

Footer: At the bottom left on the dark margin, add a tiny gray copyright-style line: “© 提示词创作分享｜用 AI 复制电影级东方美学”.

Constraints: Use exactly 3 gallery images and exactly 10 keyword tags. Keep all Chinese text crisp and legible. Avoid logos, watermarks, UI chrome, random extra sections, or additional sample images.