App内打开
歸藏: 整理了一套提示词，在 AI Studio 的 build 模式用这个提示词一键搞定：
Act as a World-Class Frontend Engineer and UI/UX Designer.
Create a "High-Fidelity Interactive Visual Novel" web application using React (es6 modules), Tailwind CSS, and the Google Gemini API (@google/genai).
### 1. Design Philosophy & UI Style (Crucial)
The UI must strictly follow a "Japanese Commercial Galgame (Visual Novel)" aesthetic.
- **Vibe:** Dark Fantasy, Cyberpunk, Polish, High-End.
- **Art Style for AI Generation:** Enforce "Japanese Celluloid Anime Style (赛洛鲁/平涂)" for all generated images.
- **Dialogue Box:** NOT a simple rectangle. It must feature:
- A "Glassmorphism" background (dark, blurred).
- Decorative Tech/Fantasy borders (double lines, glowing neon accents in Cyan/Gold).
- A dedicated, stylized "Nameplate" attached to the top-left of the box.
- A large, rectangular Character Avatar (3:4 ratio) on the left side with a decorative frame.
- **Typography:** Clean, readable serif or sans-serif with typing writer effects.
- **Animations:** Smooth fade-ins for backgrounds/characters, shake effects for intense scenes, flash effects for combat.
### 2. Core Features
1. **Story Engine:** Iterate through a predefined script (I will provide the story content below).
2. **AI Image Generation (Gemini Nano/Flash):**
- **Backgrounds:** Automatically generate scene backgrounds based on the current plot location using `gemini-2.5-flash-image`.
- **Character Tachie (Portraits):** Automatically generate character standees. Apply a "Card" style frame (rounded corners, inner shadow, border) so they look like polished assets, not raw photos.
- **Prompt Engineering:** Automatically append "Japanese Celluloid Anime Style, masterpiece, flat color" to all image prompts to ensure consistency.
3. **Text-to-Speech (TTS):** Use the Web Speech API to read dialogue. Assign different pitch/rate to different characters (e.g., Deep/Slow for the Monster, Normal for the Protagonist).
4. **Gemini Insight:** A button to analyze the current scene's sentiment and lore using `gemini-2.5-flash`.
### 3. Data Structure
- Use a constant file to store the Story Script.
- Define Characters with properties: `id`, `name`, `visualDescription` (for AI generation), `voicePitch`, `voiceRate`.
- Define Backgrounds with properties: `type`, `visualDescription`.
### 4. Story Content (The Novel)
Please parse the following novel content into the structured script.
[在此处粘贴你的小说正文内容]
### 5. Technical Requirements
- Use `process.env.API_KEY` for Gemini.
- Handle async image generation gracefully (show a "Generating..." loader in the UI).
- Single file structure preferred or standard modular structure (App, components, services).
- Ensure the layout is responsive but optimized for Landscape viewing.
Generate the complete code now.
歸藏: 最近抖音很火的即梦或者豆包直出三宫格氛围人像照片
只需要拿你的照片加上提示词就能搞定，建议用 2:3 比例
提示词在下面，整了三套不错的👇：
提示词 1：
一张以图片人像为主角的三宫格胶片质感艺术感写真图,场景为清晨安静的图书馆,阳光从高窗斜射进来。
图中人物和参考图一致,人物和脸不变,衣服为简单的白色毛衣。第一张为近景,上半身背影,人物站在高大的书架前,仰头寻找一本书,添加中英字幕“故事都写在书里吗？-Are all the stories written in books?-”第二张为中景,人物侧身坐在窗边的桌前,阳光照在翻开的书页上,低头看书,添加中英字幕“我好像…读到了别人的脚本。-I seem to be... reading someone else's script.-”第三张为大特写,人物脸部位于画面偏左侧,合上书本,眼神平静地望向窗外的光,添加中英字幕“我的故事,从这一笔开始。-My story begins with this stroke.-”整体色调清冷,带有富士胶片效果,过度曝光,画面粗粝且色调偏冷,暗部细节保留完整,高光区域呈现自然晕化、均采用柔和漫射光,无明显硬边阴影,营造出文艺且充满自我探索情绪的氛围,三张图合成一个三宫格,字幕位于底部居。
提示词 2：
一张以图片人像为主角的三宫格胶片质感艺术感写真图,场景为古典美术馆的空旷走廊,早晨或傍晚,光线透过拱形窗户洒在地板上。
图中人物和参考图一致,人物和脸不变,衣服为简约款的白色针织衫或衬衫。第一张为近景,上半身背影,人物站在一幅巨型画作前,双手插兜,背影显得修长而有艺术气息，添加中英字幕“美,是否有终点？-Does beauty have an end?-”第二张为中景,人物侧身走在长廊上,目光落在墙壁上的雕塑或另一幅画上，光影勾勒出侧脸的轮廓，优雅而富有吸引力，添加中英字幕“我只是,路过每个瞬间。-I merely, pass through every moment.-”第三张为大特写,人物脸部位于画面偏左侧,微抬下巴,眼神略带疑惑却又充满好奇,仿佛在与艺术品对话，展现出一种知性的帅气，添加中英字幕“也许我本身,就是意义。-Perhaps I myself, am the meaning.-”整体色调清冷,带有富士胶片效果,过度曝光,画面粗粝且色调偏冷,暗部细节保留完整,高光区域呈现自然晕化、均采用柔和漫射光,无明显硬边阴影,营造出文艺且充满探索与沉静的氛围,三张图合成一个三宫格,字幕位于底部居中。
提示词 3：
一张以图片人像为主角的三宫格胶片质感艺术感写真图,场景为霓虹闪烁的城市街道,刚下过雨,地面湿润反光。
图中人物和参考图一致,人物和脸不变,衣服为风衣,撑着一把透明的伞。第一张为近景,上半身背影,人物撑伞站在路口,看着对面的红绿灯和穿梭的车流,添加中英字幕“这座城市会为谁停留？-For whom does this city pause?-”第二张为中景,人物在公交站台的玻璃后,侧身看着玻璃上的雨滴,添加中英字幕“每个人都在等一趟车吗？-Is everyone just waiting for a bus?-”第三张为大特写,人物脸部位于画面偏左侧,,脸颊上有一滴分不清是雨水还是泪水的水珠,眼神平静地望向镜头外的霓虹,添加中英字幕“没关系，我的终点是我自己。-It's alright, my destination is myself.-”整体色调清冷,带有富士胶片效果,过度曝光,画面粗粝且色调偏冷,暗部细节保留完整,高光区域呈现自然晕化、均采用柔和漫射光,无明显硬边阴影,营造出文艺且充满自我探索情绪的氛围,三张图合成一个三宫格,字幕位于底部居中
