提示词

Tikdek 提示词库

浏览适用于 AI 视频生成、AI 图片生成和 API 视觉工作流的可复用提示词。

结果:1327
Product Marketing模型: seedance-2.0

Seedance 2.0 Video Prompt: Seamless Ancient Chinese Costume Change

A detailed, 10-second video prompt for Seedance 2.0, generating a sequence of seamless, rapid costume changes (Bian Lian style) for a beautiful, pure, and cute woman in various ancient Chinese styles, using smooth transitions like sleeve sweeps, hand movements, and skirt movements, all set to a tight rhythm.

Seedance 2.0 Video Prompt: Seamless Ancient Chinese Costume Change
提示词

人物:清纯,可爱,美女,冷白皮,眼角有一颗痣,身形纤细,皮肤白皙,小清新 0-1秒:素色简约襦裙,温婉少女,站定垂手,素手轻拢衣袖,清冷古风氛围感,暖调低饱和色调,柔光滤镜,中景跟拍,画面稳定 1-2秒:转身拂袖,衣袖扫镜头瞬间变装,换上飘逸白纱仙裙+白色发带,动作灵动丝滑,裙摆轻扬,古风仙气拉满 2-3秒:抬手轻遮眼,手落下瞬间变装,换上粉黛齐胸襦裙+流苏发簪,温柔甜美,回眸浅笑,镜头微推 3-4秒:轻撩裙摆,裙摆过镜变装,换上清冷蓝绿渐变古风裙+玉坠耳饰,气质清冷,动作优雅 4-5秒:低头轻拢长发,抬头瞬间变装,换上重工刺绣大袖衫+金色发钗,华贵大气,古风精致感 5-6秒:转身背对镜头,回眸刹那变装,换上侠气劲装短褐+高束发,英气飒爽,眼神凌厉,卡点利落 6-7秒:团扇轻遮面,扇面移开变装,换上红妆婚服襦裙+流苏步摇,明艳惊艳,氛围感浓烈 7-8秒:抬手绕发,落手瞬间变装,换上墨色暗纹古风裙+银饰,清冷高级,低饱和色调 8-9秒:披帛甩动遮身,披帛落下变装,换上渐变流光仙裙,仙气飘逸,光影柔和,唯美梦幻 9-10秒:转身提裙,变回素色襦裙,垂眸颔首收尾,温婉沉静,画面完整,运镜稳定流畅 整体风格:抖音爆款古风变装,动作卡点丝滑过渡,无生硬切镜,人物灵动优雅,暖调柔光滤镜,中式古风美学,画质清晰,氛围感拉满,中景跟随运镜,节奏紧凑。

Product Marketing模型: seedance-2.0

Chinese Success Seminar (Chicken Soup) Video Prompt

A highly detailed, three-shot prompt for Seedance 2.0 to generate a 15-second video mimicking the style of a frantic, brainwashing Chinese success seminar. The prompt specifies the style (shaky handheld camera, warm stage lighting, huge gold PPT text), the character (a sweating, intense middle-aged 'Master'), and three dramatic scenes involving shouting, pacing, and a climactic kneeling pose.

Chinese Success Seminar (Chicken Soup) Video Prompt
提示词

【风格】中国式成功学演讲,疯狂洗脑风,情绪炸裂,手持晃动镜头,暖色舞台聚光灯,背景是大金字PPT。 【时长】15秒 【角色】大师:中年男子,大背头油光锃亮,穿得像个新郎官(紧身西装),戴着耳麦,满头大汗,脖子上青筋暴起。 【场景】千人酒店会议厅,台下乌压压的人头。 [00:00-00:05] 镜头1:灵魂拷问(The Question)。 特写/怼脸拍。大师面目狰狞,手指几乎戳到镜头上,唾沫星子横飞。 动作:他猛地扯掉领带,瞪大眼睛看着你。 【台词/咆哮】:“告诉我!你甘心吗?你甘心这辈子就拿着三千块的死工资吗?回答我!!” [00:05-00:10] 镜头2:底层逻辑(The Logic)。 中景/舞台游走。大师在舞台上疯狂踱步,挥舞着手臂,像是在指挥千军万马。 背景:PPT上写着巨大的红字:“狼性!认知!裂变!” 【台词/嘶吼】:“穷人洗澡,富人洗脑!你为什么不成功?因为你的头脑里装的全是垃圾!把你的退路给我断了!!” [00:10-00:15] 镜头3:高潮收割(The Climax)。 全景/升格慢动作。大师猛地单膝跪地,双手高举指向天空,BGM激昂炸裂。 台下观众:疯狂起立,鼓掌,甚至有人热泪盈眶。 【台词/金句】:“今晚,让我们一起,改写命运!要想富,先脱裤(脱离舒适的裤子)!听懂掌声!!!”

作者:John
Product Marketing模型: nano-banana-pro

Nano Banana 2 Eastern Dragon Prompt

A simple, non-technical prompt for Nano Banana 2, describing a close-up shot of an Eastern Dragon sketch in a notebook, where the physical entity of the dragon emerges from the paper, except for its tail.

Nano Banana 2 Eastern Dragon Prompt
提示词

近照鏡頭,桌上一本筆記本,上面晝著一隻東方巨龍,一隻相同的龍的實體,除了尾部外全部伸出紙外變成活生生的龍 --ar 4:3

作者:TWnese
Product Marketing模型: nano-banana-pro

Weekly Outfit Design Prompt

A prompt designed for Nano Banana Pro to generate a 16:9 horizontal image showing seven different full-body images of the same female character, each representing a day of the week with a unique outfit, pose, and expression based on a user-defined theme or scenario. The AI is instructed to freely design the clothing, poses, and atmosphere, with the requirement that the seven figures are clearly the same person.

Weekly Outfit Design Prompt
提示词

严格参考图中人物 画面中出现同一个女性的7个形象,对应一周的七天 明确为同一个人, 但允许她们拥有符合主题或者场景的不同的姿态、表情、动作与穿着。 主题 / 场景: [{argument name="主题 / 场景" default=""}] 请 AI 根据该主题自由设计: 服装搭配 姿态与动作 表情与气质 整体氛围与风格 不需要统一风格, 不需要刻意对齐姿势, 不需要明确规定情绪或动作。 画面要求: 横向构图(16:9) 全身呈现 多个同一人物并排出现 整体自然、真实、有变化 背景中可以加一些符合场景和主题的手账图案 上方有个符合主题的中文标题 主体名字叫[{argument name="主体名字" default=""}],其余的标题内容你自己设计 不要分析,直接出图

作者:Dr.D
Product Marketing模型: nano-banana-pro

Nano Banana 2 Prompt Failure and Comparison

A user shares a detailed image generation prompt that nano-banana-pro 2 failed to process, despite the prompt not containing any obvious violations. The user then compares the results generated by other models (Doubao, ChatGPT, and Grok) using the same prompt.

Nano Banana 2 Prompt Failure and Comparison
提示词

woman, young, East Asian, casual, modern, comfortable, ponytail with bangs, fair skin, pale complexion, visible pores, natural imperfections, oval face, delicate refined Asian facial structure, symmetrical features, Korean K-beauty makeup style, fresh natural dewy skin, gradient lips, straight brows, innocent gaze, relaxed, rim lighting, edge glow, dramatic, cinematic lighting, dramatic rim light, film lighting, dramatic three-point lighting with warm key light from left, cooler fill light from right, strong rim light creating edge separation, high contrast duality, lighting, strong contrast, deep blacks, chiaroscuro, light and shadow, sophisticated, softbox

作者:MIKAI
Product Marketing模型: nano-banana-pro

Luxury minimalist product photography based on a reference image

A prompt for generating high-end, minimalist commercial product photography in the style of luxury brand ads, using an uploaded reference image of the core product as the base.

Luxury minimalist product photography based on a reference image
提示词

请以我上传的参考图为基础,生成一张极简主义、高质感、具有奢侈品广告风格的商业静物摄影大片。 保留参考图中【核心产品】的大致造型、比例和主要配色,让人一眼能认出是同一件产品,但请整体升级画面的质感与氛围:让它看起来像国际一线香氛、护肤品或高端电子产品的广告画面。 画面采用竖版 4:5 构图,背景为大面积干净的中性色留白,例如柔和的米白、冷灰或浅米色,营造安静、克制、昂贵的影棚空间。除核心产品和极少量必要的支撑物外,不要添加任何多余装饰,不要杂物、不摆场景,只保留极简的几何支撑物即可(如一个简洁的白色方块、圆柱或一块薄薄的透明板),并且支撑物形状要简单、低存在感,只用来托举和衬托产品。 请强烈强调产品的材质细节,让光影完全服务于材质本身: 让磨砂玻璃呈现出细腻柔和的雾面质感,陶瓷呈现出哑光或轻微釉感,高光金属呈现出干净锋利的高光边缘,液体部分则要有清晰的粘稠度和体积感。避免任何廉价塑料感、复杂花纹或花哨贴纸,整体质感要冷静、克制、昂贵。 使用专业影棚布光,整体为柔光箱的柔和漫射光,在此基础上加入精准的边缘光(rim light)勾勒产品轮廓,让轮廓线干净利落、层次清晰,阴影要柔和、干净,没有噪点和脏乱的阴影块。可以采用接近微距的镜头视角,适度靠近产品,以突出材质和细节,同时在画面四周保留充足的负空间,让产品在一片留白中「独自发声」。 整体色彩保持冷静的中性色为主,产品本身的颜色可以略微提高饱和度,用来成为画面唯一的视觉焦点。不要使用夸张的渐变背景、花哨的彩色灯光或电商促销风格,不要出现大字号价格、标签或卡通图形。 最终效果必须是一张高分辨率的商业广告大片,看起来像可以直接用于杂志封底或高端品牌官宣海报的静物大片:极简、安静、高级,核心产品在画面中心自信地存在,周围是干净的留白和精确控制的光影。

作者:VoxcatAI
Comic / Storyboard模型: seedance-2.0

Video Prompt: Original Cyber Wuxia Action Epic (Detailed Storyboard)

A highly detailed, multi-scene storyboard prompt for Seedance 2.0 to generate an 'Original Cyber Wuxia Action Epic.' It specifies 8K cinematic quality, high frame rate, and a core style of 'Hyper-speed Mechanical Martial Arts.' The prompt details two original mech characters ('Gale Blade' and 'Iron Fortress'), a cyberpunk pagoda ruins setting under heavy rain, and a 25-second high-energy fight sequence with specific camera movements, effects (welding sparks, steam, energy shockwaves), and safety/copyright optimization notes.

Video Prompt: Original Cyber Wuxia Action Epic (Detailed Storyboard)
提示词

原创赛博武侠动作巨制(Original Cyber Wuxia Action Epic)。8K电影级画质,高帧率摄影。 核心风格:超高速机械武术(Hyper-speed Mechanical Martial Arts)。强调“宗师级的见招拆招”、“残影移动”和“极致的物理碰撞”。 安全与版权优化(关键): 1. 角色完全原创:通用型设计,避开任何知名IP特征。 2. 激烈感来源:通过极快的剪辑节奏、漫天的焊接火花(Welding Sparks)、高压蒸汽喷射(High-pressure Steam)和地面震荡来体现,不出现肢解。 3. 分镜策略:采用好莱坞动作片式的多机位快速切换。 【场景】 被暴雨笼罩的赛博朋克古塔废墟顶部(Cyberpunk Pagoda Ruins Rooftop)。电闪雷鸣,地面是积水的青石板和霓虹灯倒影。 【角色设计(原创宗师机甲)】 1. “疾风刃”(Gale Blade):敏捷型。修长、哑光黑与银色相间,背负一把发光的**等离子唐刀(Plasma Katana)**。风格:快准狠的剑客。 2. “不动尊”(Iron Fortress):力量型。魁梧、重型古铜色装甲,双臂装备有**液压冲击拳套(Hydraulic Impact Gauntlets)**。风格:硬气功与重拳。 【分镜脚本 (约25秒高燃剪辑)】 [00:00-00:04] 镜头1:雨中对峙·气机锁定 机位:极稳的中景接特写(Stable Mid-shot to Macro Close-up)。 动作:暴雨倾盆。两台机甲静止对立。 细节特写:雨水打在“不动尊”炽热的肩甲上瞬间蒸发成白雾。镜头快速推向“疾风刃”的光学眼睛,光圈急剧收缩锁定目标。引擎启动的低频轰鸣声。 [00:04-00:08] 镜头2:瞬动·第一回合爆裂 机位:侧面极速跟随接快速剪辑(Side High-speed Tracking into Rapid Cuts)。 动作:“疾风刃”瞬间消失,拉出一道残影冲锋。“不动尊”重拳迎击。 激烈对撞:两台机甲在0.1秒内发生了十几次高频肢体碰撞。每一次格挡都爆出**瀑布般耀眼的金色电焊火花(Cascading Welding Sparks)**,照亮了黑暗的雨夜。 [00:08-00:12] 镜头3:凌空机动·踏浪而行 机位:低角度广角仰拍(Low-angle Wide Shot)。 动作:“疾风刃”利用反作用力腾空而起,双脚在废墟的墙面上飞速奔跑(飞檐走壁),脚下踩碎瓦片,激起一串水龙。 [00:12-00:16] 镜头4:重拳轰炸·环境破坏 机位:主观视角接冲击特写(POV into Impact Close-up)。 动作:“不动尊”预判了空中的轨迹,一记蓄力重拳轰在墙面上。 视觉奇观:整面赛博古墙在巨大的机械力量下**轰然炸裂(Explosive坍塌)**,无数砖石和钢筋碎片像炮弹一样四处飞溅。 [00:16-00:20] 镜头5:拔刀·光影剑舞 机位:环绕慢动作(Orbit Slow-mo)。 动作:“疾风刃”在坠落中拔出背后的**等离子唐刀**,刀身发出刺眼的蓝白光芒。它在空中完成了一个华丽的旋风斩。 特效:雨水被高温刀锋瞬间汽化,形成一圈白色的蒸汽圆环。 [00:20-00:25] 镜头6:终极碰撞·寸劲开天 机位:超广角定格接冲击波震荡(Ultra-wide Freeze Frame to Shockwave Shake)。 动作:“不动尊”双臂交叉格挡,“疾风刃”的等离子刀狠狠斩在上面。 结局:能量武器与重型装甲的终极对撞。 特效:以接触点为中心,爆发出一团巨大的**球形能量冲击波(Massive Energy Shockwave)**。方圆百米内的暴雨被瞬间吹飞,形成一个短暂的真空地带,地面的积水呈环形炸开。两台机甲在巨大的反作用力下同时向后滑行卸力。

作者:John
App / Web Design模型: nano-banana-pro

AI-Assisted LP Page Design with Automation Focus

This prompt uses Nano Banana and Genspark to generate various types of sales landing pages (LP). It focuses on highlighting automation and algorithmic optimization features within the design.

AI-Assisted LP Page Design with Automation Focus
提示词

AI辅助LP页面设计,Nano Banana与Genspark生成{argument name="类型" default="多种"}类型销售页面,突出{argument name="重点" default="自动化与算法优化"}

Product Marketing模型: nano-banana-pro

Apple Carving Art Prompt

A prompt for Nano Banana Pro requesting an image of an apple carved into a detailed Jiangnan water town scene (pavilions, bridges, flowing water), rendered in a 4K high-definition art museum photography style.

Apple Carving Art Prompt
提示词

一颗苹果,被雕刻成精致的亭台楼阁、小桥流水,江南水乡。艺术馆摄影,4K高清

Product Marketing模型: gpt-image-2

Futuristic Brand Visual Manual

A detailed prompt for creating a sophisticated brand visual manual that balances toy-like energy with high-end tech aesthetics.

Futuristic Brand Visual Manual
提示词

为一家名为「{argument name="品牌名称" default="任何品牌名称"}」的公司打造一套精致的多页品牌视觉手册(多张图片形式)。品牌整体气质应当充满趣味、具有未来感、色彩鲜明,并且设计感强烈,在“玩具般的活力”与“高端创意科技品牌的精致感”之间取得平衡。视觉风格需采用大胆且具有实验性的品牌语言,包括:高饱和亮色、简洁几何图形、现代字体设计、富有动感的视觉节奏,以及略带超 surreal 感的图像表达。 整体氛围应传递出:创造力、年轻感、节奏感、触感体验以及概念前卫性。在表达上,需要融合消费级产品的趣味性、时尚品牌的呈现方式,以及当代数字设计的语言体系。品牌手册中应体现: 对产品细节的极致关注(product obsession)多样化的视觉实验(visual experimentation)包装设计探索(packaging exploration)编辑风格摄影(editorial photography)沉浸式零售体验(immersive retail)跨平台品牌应用 (cross-platform brand application)最终呈现应构建一个自信、当代、统一且完整的品牌世界观一既易于理解与接近,又具备强烈的艺术指导感(artdirection)在视觉上保持高度一致性,同时展现出高端创意工作室级别的美学标准。

作者:AbleGPT
Social Media Post模型: nano-banana-pro

Adding Imperfections to AI Images for Realism

This tweet discusses the concept of adding imperfections (like flour stains) to AI-generated faces to increase texture and realism. It contrasts the performance of Seedream4_5 (better with Chinese prompts) and Nano Banana (better with English prompts) and includes a prompt fragment focused on adding facial flaws.

Adding Imperfections to AI Images for Realism
提示词

增加一些缺陷,面部瑕疵,(鼻子上和脸颊上有面粉污渍)。面部的异物感会让皮肤看起来更有纹理。

作者:张鹏
Product Marketing模型: gpt-image-2

Candid Subway Photography Style

A realistic prompt for generating a candid-style photo of a woman on a subway looking at her phone.

Candid Subway Photography Style
提示词

{argument name="场景" default="地铁上"}低头看手机的{argument name="主体" default="美丽女人"},偷拍照片。

Product Marketing模型: seedance-2.0

Seedance 2.0 Video Prompt: Cats on a Scooter

A detailed video generation prompt for Seedance 2.0, describing a scene filmed from inside a car, showing a person on a white scooter with two cats in a transparent backpack. The cats are dressed in costumes and react to a car horn.

Seedance 2.0 Video Prompt: Cats on a Scooter
提示词

从车内视角拍摄的画面:透过挡风玻璃,可以看到路上有一个人骑着白色滑板车。那人背着一只透明的背包,背包里有两只猫。其中一只猫穿着鲨鱼装,另一只则戴着蓝色兔子帽。 两只猫在背包里嬉戏打闹,突然听到汽车喇叭声后,它们同时回头,朝摄像机方向哈气。画面中,猫咪的毛发纹理清晰可见;背景是真实的城市日光场景;视频分辦率为4K,动作流畅,猫咪的表情也非常生动。

作者:John
Profile / Avatar模型: nano-banana-pro

New Chinese Style Portrait with Image Consistency Requirement

A prompt designed for generating a high-quality commercial portrait of a young East Asian woman in a red Qipao, specifically requiring the model to maintain the facial features of an uploaded reference photo. It emphasizes warm golden hour lighting, high contrast, and a framed composition using red paper-cut art bokeh.

New Chinese Style Portrait with Image Consistency Requirement
提示词

确保图像中的人物维持照片的五官 主体 (Subject): 年轻东亚女性,特写肖像。 表情:温柔地微笑着,脸部正对着相机 特征:皮肤质感细腻通透,自然清透的妆容。 妆发与配饰 (Hair, Makeup & Accessories): 发型:黑发向后梳理盘起。 头饰:头发上精心装饰着{argument name="头饰颜色" default="红色"}的中国传统剪纸艺术品,简洁大气的纹样。 首饰:佩戴一条单串珍珠项链,耳垂上有小巧的金色耳钉。 服装 (Attire): {argument name="服装款式" default="红色丝绸旗袍"}(Qipao),带有传统立领和盘扣细节。 光影 (Lighting): 暖色调侧光(黄金时刻),光源从画面左侧打入。 光线在面部和颈部投下柔和、温暖的阴影,营造出立体感和发光感。 高对比度,氛围温暖、喜庆且私密。 构图与环境 (Composition & Environment): 垂直构图 (Vertical orientation)。 前景:画面左侧被失焦(Bokeh)的较粗的红色剪纸艺术(花枝纹样)前景框架所包围,增加层次感。 背景:深红色、纯净且虚化的背景。也有一些光影效果。 浅景深,焦点锐利地集中在人物面部和眼睛。 风格与媒介 (Style & Medium): 高质量商业人像摄影,电影感光效。 新中式美学,节日氛围(春节)。 8k分辨率,超高清细节。

Product Marketing模型: seedance-2.0

Wes Anderson Style Global Travel Video Prompt for Seedance 2.0

A detailed prompt for Seedance 2.0 to generate a video of two characters traveling through six famous Chinese landmarks in the style of a Wes Anderson film. The prompt specifies clothing changes, makeup, hairstyles, camera work, music, and character actions, requiring two reference images for the characters.

Wes Anderson Style Global Travel Video Prompt for Seedance 2.0
提示词

2 人出镜(图片1 图片2,各穿不同衣服),韦斯安德森电影风格,高清画质 + 大师级运镜,画面整体黄绿色高饱和度,背景音乐为诙谐有节奏的中提琴曲,人物全程不说话按顺序打卡中国 6 大知名景点:上海→桂林→丽江→苏州→敦煌→拉萨,切换景点时二人换装当地特色女生服饰,搭配对应妆容、发型镜头快闪切入,人物脸部清晰,展现快乐笑容、灵动眼神,二人同步做行走、跳跃、摆 pose、转身面对镜头等动作;天气随机晴 / 雨 / 雪,雨雪时同撑一把伞

作者:John
Comic / Storyboard模型: nano-banana-pro

Hyper-Realistic 3x3 Grid Portrait with Fixed Identity

A complex JSON-formatted prompt designed for Nano Banana Pro to generate a 3x3 grid collage of a person, showcasing 9 different expressions, hairstyles, outfits, and backgrounds while strictly maintaining the exact facial identity and bone structure across all panels. This is ideal for creating consistent character sheets or high-quality, varied portraits from a single reference image.

Hyper-Realistic 3x3 Grid Portrait with Fixed Identity
提示词

{ "image_prompt": { "critical_instruction": "ABSOLUTE PRIORITY: Maintain identical facial structure across all panels. Do not alter underlying bone structure, nose shape, eye spacing, or jawline regardless of the expression being rendered. The identity must be unmistakable in every single shot.", "format": "3x3 grid collage", "subject": { "face_reference": "uploaded_photo", "face_identity_lock": "CRITICAL: fixed identity, zero deviation from reference facial features", "face_match_accuracy": "100% exact match enforced", "identity_preservation_details": "Ensure consistent interpupillary distance, exact nose bridge shape, and jaw structure in every panel.", "negative_constraints": [ "do not morph nose shape", "do not change eye size or spacing", "do not alter cheekbone structure", "no plastic surgery look" ], "style": "hyper-realistic portrait photography, 8k resolution, raw photo aesthetic", "skin": "highly detailed natural skin texture, visible pores, subtle imperfections, realistic complexion, consistent across all lighting conditions", "camera_settings": "shot on Sony A7R IV, 85mm portrait lens, f/1.8 aperture, sharp focus on eyes, shallow depth of field" }, "panels": [ { "expression": "joyful", "pose": "bright natural smile (maintaining jaw structure), shoulders slightly raised", "hairstyle": "high ponytail with loose strands", "outfit": "pastel cotton hoodie", "background": "soft gradient sky blue studio backdrop" }, { "expression": "surprised", "pose": "hands near face, wide eyes (without altering eye shape), natural reaction", "hairstyle": "loose natural wavy hair", "outfit": "casual cotton t-shirt", "background": "light peach studio backdrop" }, { "expression": "sad", "pose": "head tilted down, emotional soft eyes", "hairstyle": "messy low bun", "outfit": "oversized wool sweater", "background": "muted lavender tone studio backdrop" }, { "expression": "tender", "pose": "gentle smile, head slightly tilted", "hairstyle": "elegant half-up hairstyle", "outfit": "soft knit top", "background": "warm beige studio backdrop" }, { "expression": "daring", "pose": "confident gaze, chin slightly raised (showing consistent jawline)", "hairstyle": "chic slicked-back hair", "outfit": "stylish leather or denim jacket", "background": "deep teal studio backdrop" }, { "expression": "playful", "pose": "cheeks slightly puffed (ensure nose remains unchanged), playful stare", "hairstyle": "short textured bob", "outfit": "striped linen shirt", "background": "soft mint green studio backdrop" }]

第 1 / 83 页