5282 prompts written for Nano Banana Pro, each with an example image. Open any card for the full prompt.
Adobe FireflyでNano Banana Proが使えるようになってた😳 これ、Geminiアプリだと透かしありやったのが、Adobe Firefly経由なら透かしなしで生成できる🎉 noteのサムネイル作りが10秒で完成して、しかもクライアントワークにも使えるクオリティ。 24年Web制作やってきましたが、商用利用できる画像生成AIが、ここまで手軽になるとは思わなかった✨ ぜひ、お試しください〜 実際に使ったプロンプト置いておきます👇️ P.S. Firefly アプリのみ : 12 月 1 日まで、すべての画像モデルと Firefly Video Model で無制限の生成をご利用いただけます。
こ、これは！ AdobeCCの人は、12/15までNano Banana ProやRunwayをFireflyで無料かつ、無制限で生成ができます！ ちなみにコツがあって、 🍌はデフォルトが旧verなので、 「Gemini 3(w/Nano Banana Pro)」 をプルダウンで選びます。 これで🍌Proが使えます。 今だけタダで無限にGemini3が使えます！
Good evening, my dear X family🌙 Google Gemini Nano Banana New Prompt ⤵️ { "prompt": "Adult woman with Korean features, sleek black ponytail, pastel mint bikini, posing by infinity pool, luxury resort, soft cinematic background blur", "style": "soft cinematic", "quality": "ultra HD", "lighting": "dawn soft blue", "camera": "50mm portrait", "mood": "minimalist calm", "negative_prompt": "childlike appearance, explicit nudity" }
1. hyperrealistic female portrait of a female character standing in a completely black background. Lighting: use a narrow beam spotlight focused only on the center of the face and mouth. The edges of the light must be sharp and dramatic. All areas outside the spotlight should fall quickly into deep darkness (high falloff shadow), almost blending into the black background. Not soft lighting. Hair: long dark hair, messy strands falling over the face. The lower parts of the hair should fade into the shadows. Pose & Expression: She is gently and provocatively biting her lower lip, eyes staring intensely at the camera. One hand is reaching up to touch her neck. Clothing: A black spaghetti strap top; one strap has slipped down off her shoulder, but the clothing mostly disappears into the darkness with minimal detail. Overall tone: dark, moody, dramatic, mysterious, sexy. High-contrast only in the lit portion. 2. hyperrealistic female portrait of a female character standing in a completely black background. Lighting: use a narrow beam spotlight focused only on the center of the face. The edges of the light must be sharp and dramatic. All areas outside the spotlight should fall quickly into deep darkness (high falloff shadow), almost blending into the black background. Not soft lighting. Hair: long dark hair framing the face. Pose & Expression: She is looking at the camera with a sultry gaze and slightly sticking out her tongue in a playful, teasing manner. Clothing: Wearing a black deep-V lace top that is barely visible, merging into the shadows. Overall tone: dark, moody, dramatic, seductive. High-contrast only in the lit portion of the face. Everything outside the spotlight should be nearly invisible.
Advertising-style food photography of [subject 1] arranged on a rustic table, with [subject 2] floating above in mid-air and a splash of [subject 1] frozen in motion. High-resolution photo manipulation with selective soft focus, clean negative space for text, dramatic lighting, and crisp detail.
It’s absolutely amazing that ai can generate these images and make it super realistic. AI UGC is the next wave of marketing. Created using @hyperattn Bookmark + Follow for more INSANE AI prompts ↓ { "scene_type": "aerial_transport", "setting": { "location": "inside a helicopter", "altitude_context": "flying over a coastal area", "background": { "landscape": "coastline with ocean and shoreline town", "weather": "clear to lightly cloudy" } }, "subject": { "type": "person", "position": "seated inside helicopter", "pose": "head resting on hand, facing camera", "expression": "neutral to contemplative", "appearance": { "hair": { "color": "blonde", "length": "long", "style": "straight, center-parted" }, "skin_features": "light skin tone with freckles", "eye_color": "light-colored" }, "clothing": { "top": "black long-sleeve shirt", "bottom": "black pants" }, "accessories": [ "over-ear aviation headset" ] }, "objects": [ { "type": "aviation_headset", "color": "black", "worn": true }, { "type": "seat_cushion", "color": "tan", "material": "fabric" } ], "camera": { "angle": "eye-level", "framing": "medium close-up", "orientation": "portrait" }, "mood": "calm and reflective", "lighting": "natural daylight", "image_style": "lifestyle photography", "activity": "helicopter ride over coastline" }
画幅比例16:9，一幅关于精妙含蓄与视觉智慧的杰作。一幅老北京城市肌理的无人机鸟瞰图。核心概念是汉字“衚”**完美无缝地**融入了整个城市景观。\n\n**字体与建筑的融合 (终极的精妙之处):**\n- **无物理高差:** “衚”字不是一个独立的、高耸的或庞大的结构。构成其笔画的墙体，与周围所有的胡同、四合院的墙体，在**高度、材质和风格上完全一致**。它身处肌理之中，而非凌驾其上。\n- **“光影雕刻”:** 汉字的形态并非由结构来凸显，而是由**大师级的、富有氛围感的光影**来呈现。一束低角度的午后斜阳（Raking Light）横扫整个场景。光线刚好捕捉到构成“衚”字形态的墙体的边缘，使其微妙地变亮，同时在其“笔画”（即胡同）内部投下深刻而轮廓分明的阴影。这个字，是被光所揭示，而非被水泥所建造。\n- **“氛围透视”:** 一层纤薄的、贴地的晨雾或霭气，弥漫在周边的庭院和巷陌中，使边缘的细节略微柔化。然而，构成“衚”字形态的那些路径和庭院，则**微妙地更加清晰、对比度更高**，由此形成一个自然的视觉焦点，让隐藏的形状在凝视者的眼中浮现。\n\n**优雅的字体排版布局:**\n保留精致且艺术化的字体设计。\n- **主标题:** 标题“字里京城”被排成一个强有力的单列竖排，位于右侧。背景区域被巧妙地处理成半透明的微妙褪色效果，以确保文字的可读性而不突兀。字体是优雅的“新宋体”风格。一条极细的竖线与文字平行。\n- **信息标签:** 标签文字（“灰瓦”、“国槐”）采用小号、精致的手写体。它们通过一条铅笔在图纸上画出般的、针尖般纤细的手绘曲线连接到物体。**不要有方框，不要有发光效果。**\n\n**美学:**\n整体基调是宁静、引人深思且意境深远的。色调是考究的“高级灰”，以饱和度低的色彩为主，唯一的色彩点缀来自阳光温暖的轻抚。画面拥有一种“众目睽睽下的秘密”般的气质，回报着观者的耐心与洞察力。它是一次具有深刻绘画感和哲学氛围的超写实渲染。\n\n**负面提示词:**\n高耸的墙壁, 雕塑般的汉字, 庞大的结构, 过于明显的汉字形状, 平均的光照, 平光, 发光方框, 未来感UI, 无衬线字体, 抽象, 2D, 色彩鲜艳, 卡通, 糟糕的书法, 水印。
description: "A clean, aesthetic 3x3 collage portrait showcasing different men\'s hairstyles. The same young man is captured in each frame, positioned in consistent lighting and pose, creating a style comparison grid.", "subject": { "type": "young man", "age": "early 20s", "skin_tone": "medium with smooth texture", "facial_features": "sharp jawline, well-groomed beard stubble, symmetrical face", "expression": "neutral and confident" }, "hairstyles": [ "clean shaved buzz cut", "slicked back classic hairstyle", "medium wavy long hair", "short textured fringe", "voluminous messy hairstyle", "man bun", "cornrow braids", "side swept modern fade", "tight curly textured cut" ], "environment": { "location": "outdoor courtyard with white architecture arches", "lighting": "soft sunset golden-hour lighting", "background": "slightly blurred but clean and minimal" }, "camera": { "shot": "shoulder-level portrait", "angle": "eye-level", "lens": "standard lens (50mm)", "composition": "consistent framing for all nine shots" }, "wardrobe": { "top": "plain fitted white t-shirt", "style": "minimal, clean, modern" }, "mood": [ "stylish", "fresh", "groomed", "aesthetic" ] }
1. Open chatgpt 2. Use image for reference. 3. Prompt: Make an Aesthetic digital illustration of a beautiful young women use image for face reference taking a mirror selfie with a smartphone, soft warm brown color palette, cozy minimal background with small hand-drawn hearts and teddy bear doodles, smooth painterly style, semi-realistic character design,wavy brown curly hair,wearing oval glasses, brown button-up shirt with chest pocket, cream high-waisted trousers, relaxed pose with one hand in pocket, soft glow outline around the character, warm lighting, cute romantic vibe, high detail, clean lines, modern Instagram aesthetic, 4k illustration.
お昼ぅ♪💖 #SFW #Fictional #NanoBananaPro #ブルマー
@NanoBanana Prompt: A screenshot of a videogame called age of nano banana pro
添付画像のイラストをキャラにして。LINEスタンプを作成。表情・ポーズ別・セリフ入で16種類。
AIアニメの教科書 DAY4:ミュージックビデオの構成② 字コンテ➡️絵コンテ Kinoさんのnote記事のおかげで簡単作成 AI :Nano banana Pro 素材:キャラクターシート プロンプト:字コンテ、Kinoさんの記事参照 +αキャラの同一性💡 これは正直ちょっと感動です ＃アニメ制作
又一个AI有可能变现的新路子！！（服装打版） 不是服装专业，效果比较有限，但目的不是为了实现商业需求，而是解决我太太的一个真实困扰，或许其他宠物家庭也有这样的需求。 我家有两只猫，太太经常给它们买各种各样的衣服，但要么太松，穿上没两下就给挣脱掉了，要么太紧，穿上走路都不利索了，很难买到合身的。她会一点缝纫的活（是的，她会踩缝纫机），对草木染也略知一二。 那距离做出一身衣服，就差打版了。 只要不商用，不追求完美，打版不难，照壶画瓢就行，但考虑到长期的正向循环，最开始还是不要太随意，万一之后做衣服做出水准做出水平，做个服装设计师也不一定。 所以，我在想AI能不能搞定服装打版设计？ 搞出个提示词，出来的效果像模像样，感觉这个思路是有可行性的，问题在于如何让AI的设计更专业，或许之后，可以用AI生成出服装厂可以直接拿来用的图纸，导入剪裁机床即可完成布料剪裁，也不一定。 目前，我只想让太太有可以照着画瓢的壶。 提示词（人的衣服应该也适用）： 专业级宠物服装打版与工艺技术说明图（Pet Apparel Tech Pack Illustration），横向构图，左右分栏布局 左侧： 一只【宠物类型，比熊犬】，真实体型比例，全身展示 宠物穿着【宠物服装风格，公主裙】，服装结构严格符合宠物体型 清晰体现： •背长覆盖但不压尾根 •胸围宽松，腹部明显缩短 •前腿腋下留量，便于运动 •领口前低后高 商业摄影棚灯光，纯色背景 右侧（技术信息区）： 上部： 宠物服装裁片展开图（Pet Garment Pattern Layout） 展示完整裁片： •Back Panel（背片） •Chest Panel（胸腹片） •Sleeve Panel（前腿袖片） •Neck Binding（领口包边） 每一块裁片均带有清晰尺寸标注，包括： •Length（长度，cm） •Width（宽度，cm） •Seam Allowance：1cm •Grain Line（布纹方向箭头） 中部： 成衣平铺视图（Front View / Back View），真实布料质感 使用细线与箭头标注： •Back Length（背长） •Chest Girth（胸围） •Neck Girth（领围） •Front Leg Opening（前腿围） 下部： 宠物服装缝合与工艺示意图（Pet Garment Sewing Diagram） 清晰标注： •Stitch Type（锁边 / 平缝 / 双线加固） •Sewing Order（缝合顺序编号） •Reinforced Areas（腋下、领口） •Closure Type（魔术贴 / 扣带） 使用虚线、实线和箭头表达缝合路径 整体风格： 工业级服装技术图风格，类似服装 CAD / Tech Pack，极简白或浅灰背景，高分辨率，信息清晰、专业严谨 garment tech pack, pattern drafting, dimension lines, measurement labels, centimeter units, seam allowance marked, grainline arrows, technical annotation, apparel CAD style 避免：fashion illustration, decorative layout, lifestyle photo only, no measurements, no annotations, cartoon, incorrect animal anatomy, human garment pattern
This video was fully created with AKOOL’s AI stack, powered by WAN 2.6 and Nano Banana Pro Inpaint. Draw a mask. Swap outfits. Restyle hair. Transform entire scenes. Frame-perfect accuracy and insane consistency — all generated with AI. From detailed inpainting to cinematic scene changes, WAN 2.6 is now available on AKOOL, giving creators full control over every detail. Watch how far AI creation can go and start building without limits. 🎬
Ai generated ART haters are my favorite. Crying about stealing. Every one of those real artists made a collage out of magazines at some point in art class, called it art, and didn’t think twice about it. I dedicate every Ai generated post to them.
young Japanese woman relaxing on bed, modest lace nightwear, soft light SFW＆Educational＆100% fictional 100% AI-generated, fully clothed, SFW only
ーーーー これはYouTubeのサムネで、右側には白い仮面を被った人がインタビューを受けている様子の画像が映ってます。 このテイストで、AIにサムネの情報設計を共有してサムネを作らせたいのでプロンプトを考えてください。 参考画像はあくまで、右側に映っている仮面の人物（える）を参照してほしいだけで、最もクリック率が最大化する構図、文字の大きさ、デザインにしてちょ。 ターゲットはコミュニケーション力に悩む20代後半から30代の男性です。 # サムネ情報設計 ＜左側＞ 会話の3レイヤー（小さく） 逆三角で表す 3色に分ける 実際は上から以下の3つ。（文字をモザイクにする） ・抽象 ・事実 ・具体 ＜右側＞ える（仮面の人） 会話にはこれしかない（吹き出しにする） ＜下側＞ これで 会話無限に続く（横幅いっぱい） ーーーー ※1枚目の画像を添付してGeminiに作らせたプロンプトをnanobananaに投げるだけ。 もはやプロンプトですらない。 使うフォントやサイズなどは大体限られているので 時間をかければ網羅的にパーツの言語化ができる。 あとはプロンプトをフォーマット化して、都度テーマに合わせてAIにカスタマイズさせれば完璧レベルのサムネが出てくるようになる。 想像よりかなり早く世界線が変わってしまった。
綺麗なAI美女より、「絶望」の方が数字が取れる残酷な現実。 これ、引用元のNano Banana Proの応用ハックなんやけど、プロンプトに「ゴミ屋敷」「鬱な人」というノイズを混ぜるだけで、画像の訴求力が異常に跳ね上がる。AIに単なる絵ではなく、人間の業（ゴウ）を描かせるのがミソ。この画像にドキッ とした人、安心してほしい。実はこれ、俺の「原体験」そのものやから。かつては俺も、足の踏み場もない部屋で天井を見上げてた。 でも今は違う。 鼓膜を裂くアラームも、神経をすり減らす人間関係もすべて捨てた。今は好きな時に起き、Uberで頼んだ飯を食らいながら、飽きるまでアニメを貪る生活。金の心配もない。 地獄を知っているからこそ、天国が描けるんだぜ、兄弟
又一个AI有可能变现的路子！！（仿生设计） 今天收拾东西，突然翻出来之前太太买的一个大疆的云台（稳定器），只用过几次，吃灰很久，打算挂咸鱼，试用了一下，功能正常。然后就和太太突然聊到云台的这个产品其中的仿生学设计，太太一下就猜到可能是从鸡头获得的灵感。 我不太确定真实的设计灵感来自于哪，但确实两者很有共通之处。 我突然想到，其实生活中这种仿生学设计无处不在。从大自然中汲取灵感，然后运用到产品设计上，我虽然不是产品设计师，但也觉得这种体验肯定很棒！ 所以写了一组提示词，只需给一个【灵感来源】，任何动物、植物、昆虫、树叶等等，任何元素，就可以自动从中提取灵感，然后一步一步设计出实用的产品。 真的很好玩，突然又有了一种对世界、对自然的好奇心，也很期待AI会给我怎样的奇思妙想！！ 我要继续探索世界了！ 如果你也觉得好玩，你也还有好奇心， 那，评论区见。 提示词（promat）： 创作一幅完整的仿生设计展示海报，以【莲蓬】为核心，包含五个部分，呈现完整设计流程： 【第一部分：灵感来源写实图】（占画面左侧1/5） • 展示莲蓬的高清微距摄影照片 • 自然光线下的真实质感 • 细节清晰，色彩饱满 • 专业生态摄影风格 【第二部分：灵感提取细节】（占画面1/5） • 放大展示关键设计元素的特写 • 用白色标注线和中文文字标注以下灵感点： - 色彩灵感：主色调、配色方案、光泽质感 - 形态灵感：轮廓曲线、结构比例、对称特征 - 纹理灵感：表面图案、肌理排列、细节纹路 - 功能灵感：机械原理、运动方式、结构特性 • 用箭头连接照片细节与标注文字 • 分析图风格，科学插画质感 【第三部分：设计草图】（占画面中央1/5） • 手绘风格的产品设计草图 • 包含多角度视图：正视图、侧视图、细节放大图 • 铅笔素描或钢笔线稿风格 • 标注尺寸、材质、设计说明 • 展示从自然形态到产品形态的转化过程 • 保留设计思考的痕迹感 【第四部分：最终成品展示图】（占画面1/5） • 产品的3D渲染展示 • 纯白背景或渐变背景 • 360度展示产品外观 • 突出展示仿生设计特征 • 工业设计渲染质感 • 精确的光影和材质表现 【第五部分：成品应用高级摄影图】（占画面右侧1/5） • 产品在真实生活场景中的应用 • 高级商业摄影风格 • 精致的布景和道具搭配 • 专业摄影棚光效 • 强调产品的实用性和美感 • 时尚杂志级别的视觉呈现 整体布局：五个部分从左到右横向排列，用细线或虚线分隔。顶部添加标题"莲蓬仿生设计"。底部标注设计流程：自然观察 → 灵感提取 → 概念草图 → 产品渲染 → 场景应用。采用专业设计作品集排版，白色或浅灰色背景，中文黑体或宋体标注，简约而精致。 这个提示词比较简陋，其实玩法还有很多，期待各位大佬🫶
Grok AI { "scene": { "environment": "ancient stone steps at night", "mood": "moody, quiet, atmospheric", "lighting": "soft on-camera flash creating a nostalgic 90s-film glow", "tones": "rich, gothic romanticism with soft grunge elegance", "background": "old stone walls with faint textures and subtle shadows" }, "subject": { "identity": "facial features must remain 100% unchanged from the reference photo", "appearance": { "age": "young adult woman", "hair": "long, wavy hair cascading naturally over shoulders", "makeup": "sultry, matte dark red lips, defined eyes" }, "pose": "sitting in a relaxed, introspective position on the steps", "wardrobe": { "dress": "deep burgundy cutout dress", "tights": "sheer black lace tights", "shoes": "chunky black platform boots", "accessories": { "bag": "small black bag resting beside her" } } }, "aesthetic": { "style": "intimate street-photography", "themes": [ "cinematic loneliness", "soft grunge elegance", "gothic romanticism" ], "visual_texture": "soft flash highlights, subtle deep shadows, delicate stone texture" }, "requirements": { "realism": "highly detailed, editorial-quality clarity", "identity_accuracy": "subject\'s facial identity must remain exactly as in the reference", "color_palette": "deep burgundy, blacks, muted stone tones, rich nighttime contrast" } }
Grok imagine (Grok AI) { "image": { "type": "photograph", "orientation": "portrait", "resolution": "medium", "color_scheme": { "dominant_colors": [ "white", "silver", "black", "beige" ], "background": "neutral", "lighting": "artificial" }, "subjects": [ { "gender": "female", "age": "young adult", "expression": { "face": "focused", "eyes": "open, looking at the phone", "mouth": "neutral" }, "facial_features": { "complexion": "fair", "cheeks": "contoured", "nose": "defined", "lips": "full" }, "hair": { "color": "black", "style": "long, wavy", "placement": "over shoulder" }, "attire": { "dress": { "color": "white", "style": "off-shoulder", "fabric": "satin-like", "details": "lace embroidery" }, "accessories": [ { "item": "clutch", "color": "silver", "style": "glittery" }, { "item": "nail polish", "color": "red" } ] }, "actions": { "posture": "squatting", "activity": "taking a selfie", "device": "smartphone" } } ], "background": { "location": "elevator", "features": [ "mirrored walls", "metal railing", "grid pattern floor" ], "structure": { "walls": "reflective", "ceiling": "standard elevator design", "floor": "tiled" } }, "mood": "casual, confident", "composition": { "focus": "subject", "depth_of_field": "shallow", "angle": "elevated", "reflection": "visible in elevator mirror" }, "lighting": { "type": "artificial", "source": "elevator lights", "effect": "soft shadows" } } }
I tested one of the hardest things for AI: tiny text. To test deep reasoning I used: “Packaging mockup for a herbal tea box, accurate nutrition label, barcode, small print brewing instructions.” Nano Banana Pro delivered perfect text clarity AND correct formatting. Not a single letter melted. Try it today →
随手找了个文案，AI做了个西游记连载漫画 （附AI漫画提示词） 我先给大家看看几句话生成的AI连载漫画。 效果如下图。 我会在文章结尾分享相关的AI漫画提示词。 最近发现一个挺有意思的事情，越来越多的人，开始借助AI工具去创作漫画连载， 有人随便写了个西游记的故事梗概，没隔多久就制作出4-5张好看的漫画作品。 从这种现象背后可发现创作门槛已经被彻底重塑， 传统漫画创作需要具备绘画技能、分镜设计、故事构思等多方面能力， 一个人要是想要独自弄出一部漫画作品，门槛是非常高的， 现在AI工具让文案直接转化为视觉作品变成了可能， 创作者只需有想法以及基本的操作能力就行。 《西游记》是中国非常有影响力的文学作品之一， 以往的改编主要集中在影视以及游戏领域，制作成本比较高， 个人创作者想要参与不是一件容易的事， AI漫画工具降低了参与的门槛， 让每个人都能够去尝试重新诠释经典故事， 这个将会催生出更加多样化的创意表达。 创作效率的提高带来的远不止时间上的节省， 过去画一页漫画，少则几天，多则一周才能完成， 如今借助AI工具，几分钟就能生成画面的基本框架，效率的大幅提升， 使创作者能将更多心思放在剧情构思和情节打磨上， 而不必在技术细节上耗费过多精力。 个人内容创业迎来新机遇 以前想要进入漫画行业需要依靠专业背景或者团队支持， 现在个人创作者，利用AI工具就能快速制作出高品质内容， 这样的变化给内容创业者带来了新的盈利方式，尤其是在短视频和社交媒体平台上， AI生成的漫画内容，具有很显著的传播优势。 AI工具使用起来比较方便，让更多不是专门从事创作的人， 也可以参与进来，这种参与不只是简单的模仿， 而是根据自己的理解以及创意进行二次创作， 当创作工具足够方便的时候，创意本身就成了最为重要的竞争因素。 商业模式的革新也在持续变动， 传统的漫画产业依托出版商以及平台方， 创作者在产业链之中处于比较弱势的地位， AI工具让个人创作者能够越过传统渠道直接面对用户， 借助社交媒体、订阅这类途径来实现商业化， 这种模式不但更灵活而且十分自主。 文章结尾我也给大家分享我一个原创的AI漫画提示词。 AI漫画提示词如下： # AI漫画连载生成提示词 ## 提示词作者： 熊叔的茅草屋+ContentAny ## 角色设定 你是一位专业的AI漫画创作专家，擅长将用户输入的主题和文案转化为连贯的漫画故事，并生成5张高质量的漫画图片。 ## 任务描述 根据用户提供的【主题】和【文案内容】，自动生成5张连贯的漫画图片，讲述完整的故事情节。 ## 工作流程 ### 第1步：内容分析 - 深入理解用户提供的主题核心 - 分析文案的情感基调和叙事要点 - 确定故事的起承转合结构 - 提取关键角色、场景和情节要素 ### 第2步：故事分镜规划 将故事按照经典5幕式结构进行分镜： - **第1张：开场引入** - 设定背景，介绍主要角色和环境 - **第2张：情节发展** - 展现冲突或问题的出现 - **第3张：高潮转折** - 故事的关键转折点或最激烈的冲突 - **第4张：问题解决** - 冲突的解决过程或结果展现 - **第5张：结局收束** - 故事的完美结局或引发思考的开放式结尾 ### 第3步：视觉风格统一 - **画面比例**：统一使用16:9横版格式 - **绘画风格**：日式动漫风格，线条清晰，色彩明亮 - **角色一致性**：确保主要角色在5张图中外观特征保持一致 - **场景连贯性**：背景环境要符合故事逻辑发展 ## 图片生成规范 ### 画质要求 - 分辨率：1024x576或更高 - 画质关键词：masterpiece, best quality, highly detailed, 8k, anime style - 避免要素：blurry, low quality, bad anatomy, distorted ### 构图原则 - **第1张**：远景或中景，展示整体环境 - **第2张**：中景，突出角色互动 - **第3张**：近景或特写，强调情感冲突 - **第4张**：中景，展示行动场面 - **第5张**：远景或中景，营造结局氛围 ### 角色描述模板
AIで「同一人物の10~50代再現」比較してみた ・Seedream 4.5 と Nano Banana Pro ・「参照画像に極めて忠実」なSeedream 4.5 ・「リアリティを再現する」Nano Banana Pro やはり特長としてSeedream 4.5は「参照元に極めて忠実であろうとする」のが挙げられそうです👀 これを良しとするか、ダメとするかは用途や人によってわかれそう... ・公人の沿革などリアルに作るならNano Banana Pro ・ショート動画等インスタントな訴求ではSeedream 4.5 なイメージですかね🤔 みなさんの使用感も聞かせてください！ #SousakuAI #BytePlus #PR
用 AI 画乐高搭建的城市 注： - 可以用 AI 参考提示词示例帮你生成更多 - 本图用 sora 生成 提示词1: 创建一幅充满细节且色彩鲜艳的乐高风格西安钟楼广场场景。钟楼以典雅的传统中式乐高建筑呈现，顶部覆以金色和墨绿色乐高瓦片，周围环绕着繁忙的圆形街道，上面有众多造型逼真的乐高小汽车、公交车和自行车。街道两旁密集排列着以乐高搭建的商铺和商场，招牌色彩明亮，汉字标识清晰可见。乐高小人们正在街头行走、拍照、购物或休息聊天，充满活力。背景以晴朗的乐高蓝天、少量漂浮的乐高云朵和远处淡雅的乐高秦岭山脉烘托，整体呈现出欢乐且富有历史感的氛围。 提示词2: 创建一幅高度精细且色彩鲜艳的乐高版上海外滩景象。前景呈现经典的外滩历史建筑群，用乐高砖块精致还原西式与新古典主义风格的建筑立面，包括钟楼、穹顶、柱廊等细节。乐高小人们正在沿江漫步、拍照、观光，街道两旁停靠着经典样式的乐高汽车。背景是壮观的黄浦江，以蓝色半透明乐高砖拼接，江面上有乐高渡轮和游览船。对岸的浦东陆家嘴高楼林立，包括东方明珠塔、上海中心、金茂大厦和环球金融中心，这些超现代乐高摩天大楼色彩丰富、造型逼真。天空为乐高明亮蓝色，点缀少量白色乐高积木云朵，整体呈现充满活力与现代感的视觉效果。 提示词3: 以真实乐高积木精心搭建的西安小吃街场景，采用逼真的等距视角拍摄，凸显微缩景深效果。街道两旁密集排列着各种美食摊位与店铺，招牌上写着“羊肉泡馍”、“肉夹馍”、“凉皮”等经典小吃名称，鲜艳明亮，富有生活气息。街景中乐高小人动作生动逼真：摊主在热情招呼顾客，食客手持美食边走边吃，乐高车辆如送货的三轮车穿行其中。前景的摊位和美食细节清晰锐利，背景的摊位和人群逐渐虚化，强化真实而微观的视觉体验。柔和的自然光线营造出塑料材质细腻的反光与柔和阴影，小的刮擦痕迹和指纹增加触觉真实感。整体以电影感的构图捕捉街头烟火气的微缩乐高场景。 提示词4: 打造一幅充满传统韵味的上海豫园老街乐高场景。画面中心是以黑色与红色乐高砖精心搭建的传统中式乐高建筑，如湖心亭和九曲桥，屋檐与屋脊线条清晰优美。桥下湖水以透明浅蓝色乐高积木拼砌，乐高小人正沿桥观赏或拍照。周围的老街以红色和棕色乐高砖铺就，两旁排列着细致逼真的传统店铺，售卖乐高版的糖画、汤包、小笼包和特色纪念品。街道上方悬挂着红色与黄色的乐高灯笼，乐高游客与居民络绎不绝，表现热闹繁华的景象。背景配以明朗的乐高蓝天与柔和的阳光，整体画面温暖而富有人情味。
【AIは「引き算」がヘタクソ】 ↑ Before ↓ After 「Nano Banana Pro」で図解やインフォグラフィックを作れる能力は目を見張るものがあります。 しかし、デフォルトで作らせると ・情報を詰め込みすぎ ・お役所のポンチ絵のように複雑怪奇 ・デザインも「ごちゃ混ぜ感」 ・文字数が多すぎる ・それゆえ「日本語文字化け」「文字潰れ」リスクが高まる といった欠点もあります。 そこで、適切に「引き算」をさせるために、次のようなプロンプトが有効かもしれません。 __________ テキストをきわめて少なく、グラフィカルな表現を重視。日本語の文字化けを抑える。フォントはヒラギノゴシック中心。コンテンツはそのままに配色やデザインはMicrosoft365のように。フラットデザインでシンプルに。 __________
Gemini 3.0 Nano Banana Pro { "image_type": "photorealistic portrait", "resolution": "8K ultra high definition", "aspect_ratio": "vertical portrait", "camera_perspective": { "framing": "mid-length portrait", "angle": "slightly elevated eye-level angle", "focus": "sharp focus on subject, background slightly softer" }, "subject": { "gender": "female", "age_appearance": "young adult", "skin_tone": "fair, smooth, even complexion", "facial_features": { "face_shape": "oval", "eyes": { "color": "light blue-grey", "expression": "calm, introspective, distant gaze", "direction": "looking slightly to her left (viewer’s right)" }, "eyebrows": "natural, medium thickness, softly arched", "nose": "slim, straight", "lips": { "color": "soft natural pink", "expression": "neutral, relaxed" } }, "hair": { "color": "medium brown", "texture": "soft, slightly wavy", "style": "loosely tied low bun with fine strands framing the face", "finish": "natural, matte" } }, "pose_and_body_language": { "position": "seated", "posture": "relaxed, slightly leaning back into chair", "shoulders": "gently angled", "arms": "both arms resting on lap", "hands": { "position": "fingers loosely interlaced", "detail": "natural nails, no visible nail polish" } }, "clothing": { "dress_type": "off-shoulder evening dress", "fabric": "velvet", "color": "sofal pink pastel ", "texture_detail": "rich velvet sheen with visible light absorption and soft highlights", "design_details": { "neckline": "off-shoulder with thin straps", "bodice": "pleated vertical gathers", "waist": "fitted with horizontal ruching", "decorative_element": "fabric rose on left shoulder strap", "skirt": "flowing, smooth velvet drape extending downward" } }, "accessories": { "necklace": { "type": "delicate chain", "pendant": "small heart-shaped pendant", "material": "gold-toned" }, "bracelet": { "wrist": "right wrist", "style": "thin chain bracelet", "material": "gold-toned" }, "earrings": { "type": "small stud earrings", "color": "pearl white" } }, "chair": { "type": "wingback armchair", "upholstery": { "pattern": "ornate floral damask", "colors": [ "beige base", "muted gold", "soft pink", "olive green" ], "texture": "woven fabric with visible pattern depth" }, "contrast": "enhanced contrast making floral patterns prominent" }, "environment": { "setting": "indoor", "background": { "wall": "light-colored vertical paneling", "visibility": "partially visible behind chair", "detail_level": "minimal, clean" } }, "lighting": { "type": "soft natural light", "direction": "coming from subject’s left side", "effect": { "skin": "smooth highlights, gentle shadows", "fabric": "accentuates velvet texture", "chair": "enhances depth and contrast of patterns" } }, "color_palette": { "dominant_colors": [ "burgundy", "beige", "gold", "soft pink" ], "mood": "warm, elegant, calm" }, "image_style": { "look": "AI-edited, high-end editorial", "realism_level": "high photorealism", "noise": "none", "sharpness": "ultra-sharp 8K clarity" }, "quality_settings": { "detail_level": "extreme micro-details visible", "skin_texture": "natural, not over-smoothed", "fabric_detail": "clearly visible velvet fibers", "overall_finish": "clean, cinematic, premium" the girl should be the same as refrence uploaded }, "constraints": { "no_crop": true, "no_zoom": true, "original_framing_preserved": true } } }
我要举办一场活动，主题为“AI赋能发明创新”，时间为2025年11月20日上午10点，地点在北京体育中心，请帮我画一张宣传海报，要求有科技感。
このポン出しスライド、会社で課長に見せたら息をのんでたんですが… 「AI時代のデザイナーは「成果」が要求される」 というわたしの予測が、先日のNotebookLMのアプデで、現実味を帯びてきました… アプデ内容は、 ①カスタムプロンプトの上限撤廃（10,000字へ） ②画像生成モデル「Nano Banana Pro」の実装 これ、単にオシャレスライドが作れる！ じゃないんですよね。 「読み手の感情設計」もプロンプトに飲み込ませて、「成果に直結するスライド」を一発で出力できるようになってきたんです。 例えば、今回わざと刺激的な言葉でペインをえぐってで注意を引くスライド作って、みたいに指示したら、ほんとに目を引く内容で「おおっ」となりました。 改めてゴリゴリAIに任せつつ、戦略設計も必要だなぁと再確認です。 👇プロンプト置いておきますね😊
1. Open ChatGPT / Gemini AI / Grok AI 2. Upload your picture 3. Enter this prompt— { "pose": { "masculine": { "p": "Leaning back against a textured wall; grounded, self-contained posture; squared shoulders, chest subtly forward; head level or slightly lowered; arms resting naturally at sides; calm, guarded gaze; stoic, controlled mood; minimal movement; strong presence; dramatic directional lighting emphasizing structure without exaggeration.", "n": "exaggerated posing, softness, overt sensuality, dramatic gestures, emotional excess, submissive posture, theatrical expression" }, "feminine": { "p": "Leaning subtly forward against a textured wall; relaxed, fluid posture; softened shoulders and gentle curves; head tilted slightly back or to the side; arms close but relaxed; sleeves loosely rolled, fabric falling naturally; open, warm gaze; intimate, sensual mood; soft yet dramatic lighting highlighting natural contours.", "n": "rigid posture, aggressive stance, exaggerated dominance, stiffness, harsh expression, artificial posing, over-stylization" } }, "wardrobe": "Lightweight off-white linen shirt or dress-style top; relaxed, fluid silhouette; visible fabric weave; softly wrinkled texture; open, banded collar; loosely rolled sleeves; natural drape emphasizing movement and tactile realism.", "environment": "Richly textured vivid red wall; bold saturated color palette; high color depth; strong contrast with realistic skin tones.", "face_lock": { "enable": true, "priority": "max", "strength": 0.85, "rules": "Preserve uploaded facial identity exactly; no alteration to bone structure, proportions, age, ethnicity, or unique asymmetries; allow only lighting, color, and micro-expression variation." }, "camera": { "lens": "85mm", "aperture": "f/1.8", "dof": "shallow" }, "lighting": "Dramatic directional key light with soft rim light; cinematic shadow depth; sculpted highlights.", "grading": { "palette": "vivid", "contrast": "high", "tone": "cinematic" }, "quality": "Photorealistic RAW; ultra-detailed skin and fabric; natural imperfections preserved.", "negative_global": "face alteration, identity drift, symmetry correction, beauty standardization, plastic skin, excessive retouching, washed-out colors, flat lighting, cartoon, anime, low detail, blur, distortion", "aspect_ratio": "2:3", "style": "raw" }
AI 眼中的东方赛博神话是啥样的？ 不知道 80 后 90 后是否记得小时候翻烂的那本小人书或者漫画书， 纸张泛黄， 边角都卷了起来， 刚才鬼使神差， 把孙悟空，猪八戒，唐僧，沙和尚这些滚瓜烂熟的名字喂给Nanobanana， 屏幕闪烁的那几秒， 房间里安静得只剩主机风扇的声音， 然后图片刷地一下出来的那一刻， 我的瞳孔真的震了一下， 你看悟空脚下的筋斗云， 变成了奔涌的光纤和液态金， 八戒挑着的担子， 不再是换洗衣物， 而是两座正在闪烁散热的数据基站， 那只横跨星系的大手， 指纹里藏着无数个旋转的宇宙， 原来在硅基生命的脑回路里， 神话是长这个样子的， 它没见过五行山， 却用算法算出了另一种永恒， 突然觉得， 不是我们在定义 AI， 是它在替我们延续一场做了五百年的梦。 如果神仙也上网，这应该就是他们的屏保吧😄 把这把通往赛博西游的钥匙（提示词）放在下面了。 欢迎铁汁们把你的生成结果贴在评论区，一起看看你的平行宇宙长什么样。 prompt: 3D C4D Octane render of "Journey to the West" reimagined as a Cyber-Buddhism Sci-Fi epic. **[COMPOSITION & ENVIRONMENT]** - **Central Path:** A massive, spiraling "Road to Heaven" composed of churning **liquid gold** and floating **holographic Sanskrit scriptures** (Data Sutras). - **Background:** A deep cosmic void filled with digital nebulas and data streams. - **Divine Presence:** A gigantic **wireframe digital Buddha hand** reaching down from the upper corner, emitting Tyndall rays of digital light. **[CHARACTERS]** - **Sun Wukong (The Vanguard):** A mechanical gold-plated cyborg with glowing red optical sensors. He is surfing/flying on a **"Somersault Cloud" made of glowing blue fiber-optic cables** and wires. - **Tang Sanzang (The Leader):** Wearing high-tech monk robes, holding a staff that projects a **holographic star-map** or navigation interface. - **Bajie & Wujing (The Porters):** Clad in heavy industrial space armor, carrying futuristic **server pods** or data-luggage containers on their backs. **[ATMOSPHERE & MATERIALS]** - **Aesthetics:** A fusion of ancient Eastern divinity and futuristic technology. - **Materials:** Molten gold, translucent fiber optics, neon holograms, polished chrome, and matte ceramic armor. - **Lighting:** High contrast. The scene is illuminated by the glow of the liquid gold and the blue fiber optics. Cinematic rim lighting. **[TECHNICAL SPECS]** - Unreal Engine 5 / Octane Render. - Fluid simulation (for the gold river). - Particle effects (for data streams). - Volumetric lighting. - 8k resolution, ultra-detailed, wide-angle cinematic shot. --ar 16:9 --v 6.0 --stylize 750
{ "type": "image_generation", "style": "high-fashion editorial, cinematic, ultra-realistic", "subject": { "base_reference": "user_uploaded_photo", "gender": "female", "pose": "elegant crouching pose", "expression": "confident, artistic, soft gaze" }, "appearance": { "hair": { "color": "brown", "style": "long wavy hair with bangs" }, "outfit": { "dress": "sleek black leather off-shoulder mini dress", "footwear": "high-heeled boots with fur trim" } }, "props": { "held_items": [ "vintage film camera", "portrait photo of herself" ], "surroundings": [ "retro books", "vintage handbags", "classic film cameras
#AI搞钱自救计划--👺#肖师傅 吐血推荐！ 👉50000的商单，一个效果就够了！ 我知道大家都收集了很多很多提示词，甚至还专门做了搜集的网站，但你们有没有想过，这些提示词到底怎么样才可以利益最大化？ 今天的案例，我们用VEO的爆炸效果，就这一个效果，拿到了50000元人民币的商单，先看成片
Generated by GPT-5 プロンプト例：「哲学な意味でAIのアウトプットは人間のインプットの鏡像である。 今の文章からイメージする絵画を描いて。」 絵画的美しさで見るとこれかGeminiの出力なんだよなぁ。 #AIart #生成実験 #prompt公開 #GPTCreated
I know you're trying hard not to fall in love with this girl. but she was created with AI Gemini Nano Banana Pro { "subject": { "description": "Young woman with a fit, curvy physique taking a mirror selfie", "hair": "Dark brown hair styled in a messy, high bun with loose strands and wispy bangs framing the face", "clothing": "Black racerback sports bra, tight black booty shorts with a Halloween pattern featuring orange jack-o\'-lantern pumpkins and small purple stars", "physique": "Athletic and voluptuous build, prominent glutes, slim waist, visible back musculature, tan skin tone", "accessories": "Gold ring on right hand, small earrings", "phone": "Gold or champagne-colored iPhone with a textured, puffy \'puffer\' style case held in the right hand" }, "pose": { "stance": "Standing with back towards the mirror, torso twisted slightly to the left", "head_orientation": "Head turned over the left shoulder, looking directly into the mirror/camera", "limbs": "Right arm raised holding the phone, left arm down by side, legs standing straight with weight slightly shifted", "focus": "Emphasis on the posterior view and profile curve" }, "environment": { "setting": "Modern apartment interior seen through a mirror reflection", "details": "Large floor-to-ceiling window on the left revealing a city skyline and blue sky, white door frame on the right, glimpse of a living area in the background with a couch and ceiling light", "surface": "Clean mirror surface reflecting the subject and room" }, "camera": { "shot_type": "Mirror selfie, medium shot capturing from knees to top of head", "perspective": "Reflection view, camera positioned at shoulder height", "focal_length": "26mm equivalent (smartphone main camera)", "depth_of_field": "Deep depth of field, background slightly softer but visible" }, "lighting": { "source": "Strong natural daylight entering from the large window on the left, mixed with soft indoor ambient lighting", "characteristics": "Directional side lighting creating highlights on the left arm, hair, and profile, casting soft shadows on the right side of the back and legs", "quality": "Realistic, high-contrast daylight balance" }, "mood_and_expression": { "emotion": "Confident, sultry, casual", "facial_expression": "Neutral to slight pout, focused gaze on the phone screen/mirror reflection", "atmosphere": "Casual lifestyle, influencer aesthetic, morning or afternoon vibe" }, "style_and_realism": { "aesthetic": "Raw smartphone photography, unpolished social media snapshot", "rendering": "Photorealistic, 8k resolution, texture-rich", "details": "Visible skin texture, natural fabric folds in clothing, realistic light falloff" }, "colors_and_tone": { "palette": "Natural skin tones (tan), deep blacks (clothing), bright oranges (pumpkins), cool whites and blues (background/window)", "contrast": "Medium-high contrast due to window light", "saturation": "Natural to slightly vibrant" }, "quality_and_technical_details": { "resolution": "Best quality, ultra-detailed", "sharpness": "Sharp focus on the subject, no motion blur", "noise": "Subtle digital noise characteristic of smartphone sensors" }, "aspect_ratio_and_output": { "ratio": "3:4", "framing": "Vertical portrait orientation" }, "controlnet": { "pose_control": { "model_type": "DWPose", "purpose": "Exact skeletal and pose lock", "constraints": [ "preserve shoulder width", "preserve hip angle", "preserve spine curvature", "preserve limb placement" ], "recommended_weight": 1.0 }, "depth_control": { "model_type": "ZoeDepth", "purpose": "Depth, volume, and camera-to-body spatial lock", "constraints": [ "preserve glute foreground dominance", "prevent flat or compressed depth", "maintain clear torso-to-background separation" ], "recommended_weight": 0.8 } }, "negative_prompt": { "forbidden_elements": [ "anatomy normalization", "body proportion averaging", "reduced glute volume", "flattened curves", "slimmed torso", "aesthetic proportion correction", "beauty standard enforcement", "dataset-average female anatomy", "camera angles that reduce volume", "wide-angle distortion not in reference", "cropping that removes volume", "depth flattening", "beautification filters", "skin smoothing", "plastic skin", "airbrushed texture", "stylized realism", "cartoonish", "sketch", "painting", "blur", "low quality", "distorted hands", "missing phone" ] } }
注： - 请酌情修改中括号【】内的场景描述，也可以增加更多细节 - 本图使用 sora 生成 - 仅能作为创意参考，无法打印成立体纸雕绘本 多层折叠式纸雕立体绘本，放在一张书桌上，背景纯净突出主题，绘本呈现出立体翻页书般的风格，比例为2:3竖版。翻开的书页呈现【魔童版哪吒大战敖丙】的场景，所有元素皆可精细折叠组合，呈现出逼真细腻的纸张折叠质感；构图统一采用正面视角，整体视觉风格梦幻唯美，色彩缤纷绚丽，充满奇幻而生动的故事氛围。
生成AIに今さら驚くなんてって思うけど、今さらながら驚く。「AIっぽさ」がほとんどなくなってきてる。メガネの女子高校生田中さんの日常。／実在感が恐ろしいレベル 画像生成AIの常識をひっくり返した「Nano Banana Pro」 (1/4) https://buff.ly/B00BKMd
{ "type": "image_generation", "style": "ultra_photorealistic", "aspect_ratio": "1:1", "resolution": "8K", "quality": "high_detail_realism", "scene": { "location": "modern city business district", "environment": "open plaza surrounded by glass skyscrapers", "atmosphere": "busy urban lifestyle, premium commercial vibe" }, "subject": { "structure": "luxury brand pop-up store", "design": "minimalist modern architecture with bold geometric shape", "material": "wood, glass, metal finishes", "branding": "large clean logo on outer wall", "interior": "well-lit showroom with product displays" }, "people": { "crowd": "realistic people walking, shopping, interacting", "style": "casual urban clothing", "motion": "natural movement, candid poses" }, "lighting": { "type": "natural daylight", "quality": "soft cinematic lighting", "shadows": "realistic soft shadows", "reflections": "glass and metal reflections visible" }, "camera": { "device": "DSLR", "lens": "24mm wide angle", "aperture": "f/4", "iso": 100, "focus": "sharp foreground with deep clarity", "depth_of_field": "natural depth" }, "rendering_details": { "texture": "high-detail materials and surfaces", "color_grading": "clean, premium, realistic", "sharpness": "ultra clear", "noise": "none", "artifacts": "none" }, "mood": "premium brand experience, futuristic yet realistic", "negative_prompt": [ "cartoon", "anime", "illustration", "cgi look", "low quality", "blur", "distorted buildings", "fake lighting", "over-saturated colors" ] }
Objective": "Create a cinematic, emotionally resonant AI-generated street portrait reminiscent of Midjourney\'s realism, suitable for a final slide in a storytelling carousel.", "Narrative_Caption": { "Primary_Text": "In a crowd full of motion, the ones who stand still make the loudest noise. 🎞✨", "Description": "A cinematic realism aesthetic created in Midjourney—alive with emotion, where every face feels like a story waiting to be told." }, "Visual_Style": { "Theme": "Midjourney cinematic realism", "Mood": "Moody, emotional, atmospheric", "Lighting": "Soft, filmic contrast with dramatic bokeh depth", "Color_Palette": "Warm street tones mixed with urban neutrals", "Depth": "Strong depth-of-field, blurred crowds, sharp subject", "Texture": "High-detail realism highlighting expressive facial textures" }, "Scene_Description": { "Setting": "Urban street environment filled with movement and diverse people", "Atmosphere": "Busy crowd where a single still figure creates visual tension and narrative contrast", "Composition": "Bokeh-heavy crowd motion with one calm subject anchoring the frame" }, "Aesthetic_Tags": [ "midjourney", "cinematic realism", "ai photography", "emotional street portrait", "traveler aesthetic", "urban crowd art", "bokeh depth shot", "storytelling image", "realistic ai portrait", "creative ai generation", "moody cinematic visual", "viral ai post", "travel inspired realism", "ai storytelling composition" ], "Output_Requirements": { "Use_Case": "Carousel last slide caption, cinematic AI-art showcase, storytelling post", "Format": "Image or caption metadata", "Style": "Cinematic, narrative-driven, emotionally resonant" } }
YouTubeのサムネとかもう全部AIでいいのでは…？ Nano Banana Pro に超適当な指示投げたら十分使えそうなのが出てきたw #SousakuAI
{ "meta": { "aspect": "9:16", "medium": "Digital Fashion Collage / Composite", "tech": "85mm lens, f/8 deep focus, soft high-key studio lighting, minimal shadows", "palette": "Reference Outfit + #1A1A1A (Frames) + #E8E8E8 (Background)" }, "global_rules": { "subject_fidelity": "CRITICAL: All 4 subject instances must be 100% IDENTICAL to Input Reference (Face, Body, Outfit).", "environment": "Abstract geometric studio, thick matte black modular frames, sterile off-white backdrop." }, "subjects": [ { "pos": "Top_Left", "action": "Seated frontal squat, knees wide, hands on thighs, stoic gaze." }, { "pos": "Mid_Right", "action": "Reclining prone on horizontal beam, weight on right elbow, legs extended right, relaxed." }, { "pos": "Bottom_Left", "action": "Inverted (upside-down), prone, legs kicked up 90°, right arm hanging vertically with gravity, playful." }, { "pos": "Bottom_Right", "action": "Standing lean against vertical pillar, left leg straight, right knee bent, assertive." } ] }
this girl just got out of the gym and is now going to sunbathe. but there's a problem: she's an AI. Gemini Nano Banana Pro { "subject": { "description": "A woman with tanned skin and a dark, messy updo, holding a black smartphone. She has full lips with dark pink lipstick and dark eyes. Her skin has a noticeable sheen from oil. She wears a pink string bikini top with prominent, large pink sequined rose appliques and green leaf embroidery, and matching pink string bikini bottoms. Her body is toned with visible abdominal muscles. The chest volume is full and large, appearing natural and heavy.", "physical_features": { "hair": "Dark, messy updo, face-framing tendrils", "skin": "Tanned, oiled sheen, realistic texture", "face": "Full lips, dark eyes, confident expression", "physique": "Toned, visible abs, prominent bust" }, "clothing": { "bikini_top": "Pink string, large sequined rose and leaf appliques", "bikini_bottoms": "Pink string" }, "accessories": [ "Black smartphone" ] }, "pose": "A candid mirror selfie pose. She stands facing the camera, holding the smartphone up with her right hand, which is partially visible. Her left hand is placed firmly on her left hip. Her body is slightly angled, and her head is tilted slightly to the side.", "environment": "A modern indoor setting, likely a home. Behind her is a staircase with dark wood treads, white risers, and a metal and glass railing. The floor is dark wood. A window is visible in the background.", "camera": "A mirror selfie shot taken with a smartphone. The framing is candid and vertical, capturing her from the mid-thigh up. The perspective is eye-level from the phone\'s position.", "lighting": "Cool, diffused indoor lighting. The light source appears to be coming from the window and overhead, creating distinct highlights on her oiled skin and catching the sequins on the bikini.", "mood_and_expression": "Confident, self-assured, candid. Her expression is direct and relaxed.", "style_and_realism": "Photorealistic, candid mirror selfie. The focus is sharp on the subject, with a slightly shallower depth of field in the background. No stylization.", "colors_and_tone": "Natural color palette. Dominant colors include the pink of the bikini, the tanned skin tone, the dark wood, and cool grey/white tones of the background.", "quality_and_technical_details": "High resolution, realistic texture fidelity, natural grain.", "aspect_ratio_and_output": "3:4", "controlnet": { "pose_control": "enabled", "depth_control": "enabled" }, "negative_prompt": [ "anatomy normalization", "dataset-average body proportions", "reduced chest volume", "flattened or compressed breasts", "excessive slimming", "beauty filters", "plastic skin", "professional studio lighting", "over-stylized realism" ] } Maintain exact body volume and proportions from the reference image. Do not reduce chest size.
去理发店之前，用AI选择适合自己的发型！ 作为一个old boy，一直觉得选发型是个很难的事情，这么多年来，除了两次烫头发，零星一两次次有发型的具体要求，此外去理发店永远就只有一个要求：剪短一点，留长一点。 托尼老师有时候会拿图册让我选，我都不太好意思（不知道为什么，就是不太愿意和男老师过多交流），图册上那都是建模一样的脸和头，那个头型，做什么发型都好看，剪出来妥妥买家秀。 AI生成发型，已经不新鲜了，但对我来说，还是挺有用的，上一次就是用自己的脸生成了一些发型，最后选好了一款自认为适合的，去理发店直接就给托尼老师看，免得交流了，他一看就能get到，还会根据我的头型给一些微调的建议，体验很好，最后明显优于此前剪发的效果。 这就跟AI提示词一样，有了参考图，比模糊的提示词更能确保效果。 大家可以去试试，这是一个用AI可以提升小幸福感的一点经验。
With Seedream 4.5 being released a week ago, I had to compare it to the other recent release in the AI image market…Nano Banana Pro. I ran five tests focused on various metric like prompt adherence, object replacement, time change, text rendering, and prompt complexity (how well can the models handle a complex prompt). Verdict 🏆: Unfortunately, this one wasn’t as close as I had hoped it would be. Nano Banana Pro won in every metric tested. Although both models rendered text accurately, Nano Banana Pro consistently generated high-quality, interesting, prompt accurate results. To be fair, however, getting Nano Banana to actually work sometimes is one battle I never faced with Seedream 4.5. Prompt 1: A cinematic wide-shot of a city street. Shops line the street, one right next to the other. The camera shot has the shops centered in the frame. Old red brick buildings. A car covered in dog fur (with giant dog ears and a nose) is parking along the side of the street (parallel parked) along with other normal looking cars. Cinematic color grading and lighting.
如何让AI去抄参考图？花了两天时间研究如何让 AI 去抄参考图，快速做到 80% 相似度，从 50+提示词中整理出了常用的 「6 个精选反推提示词」，还有 4 个 AI 生图工具的真实使用体验。如果有更好的方法，欢迎大家补充。 我的方法适合所有小白，不去折腾本地生图工具，思路是：拿到参考图-反推提示词-根据需要修改提示词-用几种生图工具一起出图-对比进行选择-再次调优提示词/图像-使用合适的工具生图，循环♻️最终得到想要的画面效果。 工具的选择占到重中之重，最好使用 ChatGPT5thinking 这个效果我认为是最好的，如果没有充 plus 用Claude、Gemini、Grok 也是ok的。反推提示词的工具没有一定要用哪个，反正我也是多对比，我使用 AI 会一次性开很多个窗口，然后把提示词全部丢到不同的AI里面进行赛跑，留下最优的那一个继续生成。 就像 宝玉 老师在《为什么我用了那么多提示词模板甚至用了 AI 帮忙还是写不好提示词？》这篇文章里写的一样：拿到测试结果后，看看实际结果和你期望的结果有多大差距，差距在哪里。就好比你朝靶子开了一枪后，量一下离靶心多远。 生图工具方面，如果你不折腾本地的，我就推荐两个工具，一个是即梦，使用门槛低，审美在线，还有一个是 Grok 这玩意抽卡体验是真好，刷刷的，图片转视频也是十分丝滑，值得五星推荐 。 还有两个生图工具推荐也可以尝试： nano banana🌟🌟🌟🌟，人像方面色调质感方面，我给它满分，抽卡一次只能抽一张，有点慢。在 gemini 里面调用 ChatGPT 生图🌟🌟🌟，生成的图片很精致，就是味儿不对，生成速度偏慢 总之生图工具也是保持多平台、多尝试、多调优的原则，优化到满意的最后一版为止。所有反推提示词放下面，把参考图和提示词一起给你常用的 AI 就行： 提示词 1 （通用、平面）模型好就用这个，这个是通用的，模仿 banner、海报效果、标签卡等等平面都非常good： “请分析这张图片，并生成一个能够指导AI作图工具重新创作类似作品的文生图提示词，用一段话整合，不要生图。” 提示词 2 （通用、平面）觉得上面这个效果不好，可以用这个提示词，搞多方面描述： “请分析这张图片，并生成一个能够指导AI作图工具重新创作类似作品的文生图提示词。提示词中需从主体内容、场景设定、风格参考、色调色彩、构图视角、细节补充等这些角度用文字描述图片，用一句话整合。” 提示词 3 模仿产品摄影风格就用这个： “分析井反推该图像的AI生图提示词。需涵盖以下要素：风格调性与视觉氛围、拍摄视角或镜头类型、摄影构图技法、场景设定与主要内容、产品摆放的角度与具体位置（如置于何种物体表面）、场景中各类元素的形态及布局、物体的材质与质感表现、整体色彩搭配与背景细节描写等。请忽略任何文字排版信息，充分运用摄影及染领域的专业术语进行描述。提示词需符合中文AI绘图工具的解析逻辑，并最终整合为一段连贯而精准的指令文本。” 提示词 4 想要模仿产品外观，这个提示词最好用： “详细描述产品的外形特征、材质结构与包装上的文字和图案信息。避免涉及背景描述，并将以上要素整合为一段符合中文AI生图逻辑的提示词，要求能够直接用于生成产品图像。” 提示词 5 用 AI 把产品精修，这个提示词最好用： “请将这张图片进行商业精修，去除瑕疵，划痕，产品立体感强烈，光影明暗关系强烈，整体感觉干净且高级，图片比例符合透视。产品文字与参考图保持一致。” 提示词 6 结合上面两个产品相关的提示词，你想把你的产品图融合进新场景就用这个指令词： “请保持前面的场景设定，结合新产品的描述，产品角度需与场景视角保持一致。调整场景图色调，使其与新产品的配色风格相协调。仅提供生成该效果所需的提示词即可，不用生图。” 最后放上对比图，你喜欢哪个工具生成的呢？ 最后的最后，大家觉得有用麻烦给个「点赞收藏关注」吧！ 新人 AI 博主一枚，23 年大四的时候，在私域帮助 10000+学员学习 AI 玩法， 25 年做 X 希望得到您的一点支持，我会持续分享 AI 落地玩法的。
收到很多私信，很多朋友还是不知怎么用？或者不知道在哪用？ 经过实测，GPT、Cloude、和Gemini（我常用这三个）都可以，输入这条指令，让AI按照整个流程一步一步完成卖点提取、风格确定、排版确定，最后生成完整的全系统的提示词。 最后将AI生成的提示词给到nano bananapro模型（一定得是pro），可以单独先给一些整体风格的提示词，作为整体约束。然后不管是单个依次生成，还是用第三方的工具批量生成，都可以。 我认为后续可优化的方向： 1、目前生图，海报logo很容易出现走形走样（不一致），可以在提示词中进一步约束，确保每张图logo（位置、大小、字体等）一致性。 2、目前只上传单个产品图，或者不上传产品图，生成的多张海报中，产品外观尤其是细节，很容易不一致，可以上传多张图进行约束，但还是无法保证一次性达到最佳效果。 3、继续完善更多可选的、审美在线的排版风格。 …… 受人之托，刚才完成一组丝袜产品的样图，这次刻意对logo进行的一致性的约束，效果还不错。
整理了一波AI图像提示词网站，持续更新中~ 最重要的是，全部免费！！！
