热门角色不仅是灵感来源,更是你的效率助手。通过精挑细选的角色提示词,你可以快速生成高质量内容、提升创作灵感,并找到最契合你需求的解决方案。让创作更轻松,让价值更直接!
我们根据不同用户需求,持续更新角色库,让你总能找到合适的灵感入口。
设计一个基于挑战任务的场景创意提示词,适用于艺术创意。
以下是一份“创意简报转图像方案”的提示词编写挑战。它兼具叙事性与可执行性,旨在训练你用Stable Diffusion(建议SDXL 1.0+Refiner)将文字意象化为高质量图像。请按步骤完成,并在每一步提供你的思考过程与最终提示词版本。 挑战标题:蓝时雨巷的流光茶摊 创意简报(沉浸式叙述) - 城市刚刚被雨水抹上了蓝时的微冷光泽。巷口霓虹像被潮汐浸泡的水母,柔软却夺目,色彩在湿润的地面上拉出长长的呼吸。移动茶摊停在一盏孤灯下,铜壶翻涌,蒸汽像温驯的兽吐息,带着桂花与乌龙的香气,缓缓攀上伞檐。玻璃与不锈钢上细密的雨珠折射出微型星海。两位路人擦肩而过,半透明的伞沿勾出冷暖对比的弧,脚下溅起的水花是微型焰火。远处列车划出一线光轨,如同时间的琴弦。整幅画面应该在湿度与温度的张力中呼吸:冷是蓝、暖是金,静是雾、动是光。 任务结构与交付物 1) 解析与提炼 - 从简报中提取核心视觉要素,分为:主体、场景、光影、材质/细节、情绪/叙事、色彩、镜头语言/构图。 - 输出要素清单(不超过12条),每条后附简短理由。 2) 写出基础正向提示词(Base Prompt) - 目标模型:SDXL 1.0 Base。语言建议中文或英文均可,避免堆叠无效形容词,保证每个片段有“可见性”。 - 覆盖维度:主体+环境+构图+镜头参数+光源位置与性质+材质与微观细节+氛围/天气+色彩计划+叙事动势+质量控制(如highly detailed, photorealistic/cinematic, physically based rendering)。 - 要求使用逗号分隔短语,允许使用权重语法如(关键词:1.2)或多重括号以加强元素。避免使用在地化不通用的口号式标签。 3) 负向提示词(Negative Prompt) - 目标:抑制低质量与常见瑕疵(如多余肢体、错位手指、过度锐化、涂抹感、文字水印、AI噪点)。 - 覆盖:构图错误(裁切脸/主体缺失)、解剖与手部、质感(塑料感皮肤)、影像缺陷(色带、伪影)、风格污染(卡通化或过分插画化,当你选择写实向时)。 4) 三种艺术取向分支(同一简报,三种不同落地风格) - 真实摄影向:如“蓝时城市纪实摄影” - 油画/厚涂向:可突出笔触、颜料厚度、光泽与肌理 - 平面海报/图形向:强调几何构成、负形空间、极简配色、文字留白(禁止直接生成文字,但为排版留白) - 为每个取向分别写一套正负向提示词,并说明为了达成该风格你在镜头、光线、材质、配色、构图标签上的取舍。 5) 参数与流程建议(给出你的推荐配置) - 分辨率与比例:基底分辨率建议1024×1024(SDXL原生友好),并按交付物扩展三种比例: - 竖版封面:1024×1365(约3:4) - 横版主视觉:1536×864(约16:9) - 正方形社媒:1024×1024 - 采样与CFG:DPM++ 2M Karras 或同类稳定采样器;Base 28–36 steps;CFG 5–7(SDXL常用区间)。 - Refiner衔接:在Base采样进度约0.75–0.85处切换至SDXL Refiner补光与细节(或以img2img方式,denoise strength 0.2–0.35),Refiner 8–12 steps。 - 批次与种子:每取向设置固定seed,batch 4–8,选优;记录种子便于复现。 - 放大与细化:若需更高分辨率,先以latent upscaler或高分辨率修复(1.5–2倍),再使用基于ESRGAN或其他超分模型做轻度面向细节的放大。避免过度锐化导致的边缘光晕。 6) 结构化提示词模板(供你填充) - 正向模板: - 主体与场景:… - 构图与镜头:镜头焦段、光圈、机位高度、取景比例、景深 - 光线与氛围:主光/辅光/背光、体积雾、降雨、反射 - 材质与细节:表面粗糙度、微水珠、蒸汽、金属/玻璃高光 - 色彩与对比:主色/辅色、对比关系、色温 - 叙事动势:动作/时间线索(光轨/水花) - 质量与风格约束:photorealistic/cinematic/painterly/graphic poster, high detail, clean composition - 负向模板: - 质量缺陷、构图问题、解剖与手部、风格污染、文字/水印、压缩伪影 7) 可选控制与扩展 - ControlNet建议: - Depth/Normal:巩固空间层次(建筑与小车体量) - Lineart/Scribble:用草图锁定构图与人群位置 - Reference Only/Style Reference:做色彩/光感对照(不需引用具体艺术家) - LoRA使用: - 只在明确目标风格时使用,权重在0.6–0.9间微调,避免风格盖过叙事元素。 - 多轮迭代: - 第1轮定位构图与光比,第2轮校准材质与色温,第3轮抛光细节与降噪。 评分标准(自评或团队评审) - 可视化忠实度:是否还原“冷暖对比、湿度与温度的张力、光轨+水花”的叙事核心 - 风格一致性:三种取向各自清晰且不混淆 - 可复现性:参数、种子与流程可复刻;提示词无冗余噪声词 - 画面控制力:主体突出、构图稳定、细节可信、无明显AI瑕疵 示例答案片段(供参考,可直接替换你的元素) - 真实摄影向 正向(英文示例,便于通用性): blue-hour rain-soaked alley in a dense city, mobile tea cart with steaming copper kettle, porcelain cups, bioluminescent-like neon reflections on wet asphalt, two pedestrians crossing with translucent umbrellas, distant train light trail, rule of thirds, leading lines from alley edges, low angle eye-level, 35mm lens, f/1.8 shallow depth of field, soft backlight + practical neon key light, volumetric mist, fine rain droplets, micro water beads on stainless steel and glass, glossy specular highlights, cool blue ambience with warm golden steam glow, cinematic, photorealistic, highly detailed, clean composition - 真实摄影向 负向: lowres, oversharpened, watercolor, cartoon, extra fingers, deformed hands, duplicate limbs, bad anatomy, text, watermark, logo, color banding, blown highlights, underexposed shadows, jpeg artifacts, tilt shift miniature effect - 油画/厚涂向 正向(中文示例): 蓝时雨巷中的移动茶摊,铜壶蒸汽如温暖吐息,雨珠在不锈钢与玻璃上形成颗粒高光,霓虹在湿地上映出长条冷色反射,两名撑伞路人掠过,远处光轨划破暗蓝空气;厚涂笔触可见,边缘略带笔锋毛边,颜料堆积在高光处反射出油性光泽;三分法构图,前景水花、中景茶摊、远景光轨,冷蓝主调配金色点亮,戏剧化对比,具象却保留笔触肌理,painterly, impasto texture, rich brushwork, controlled edges, gallery quality - 油画/厚涂向 负向: photorealistic skin, plastic shine, posterized gradients, muddiness, excessive blur, low contrast, text, watermark - 平面海报/图形向 正向: minimal poster of a neon-lit rainy alley, simplified geometric shapes, bold silhouette of a mobile tea cart with rising steam, glossy puddle reflecting cyan and deep blue blocks with a small wedge of warm amber, two umbrella arcs as clean curves, wide negative space for typography, strong diagonals, high contrast, flat shading with subtle grain, crisp edges, graphic poster, Swiss grid influence, print-ready, vector-like clarity - 平面海报/图形向 负向: photorealistic textures, noisy film grain, painterly strokes, complex clutter, text, watermark 推荐参数(可按实际环境调整) - 模型:SDXL 1.0 Base + SDXL 1.0 Refiner - 采样器:DPM++ 2M Karras - 步数:Base 32;Refiner 10(切换于0.8进度或以denoise 0.25在img2img执行) - CFG:6 - 分辨率:1024×1024起步;按目标延展为3:4、16:9与1:1 - 批次与种子:batch 6,固定seed记录;选择最佳图进一步放大与精修 提交清单 - 要素提炼清单(≤12条) - 三套完整提示词(每套含正负向) - 参数配置与流程说明(含种子) - 3张不同取向的成品图(若无法出图,请提交你的提示词与预期图像描述) - 迭代记录:至少2轮修改点与改进理由 创作提醒 - 用“看得见”的词:能在画面中被明确渲染的实体、光位、材质与动作 - 控强弱:用括号权重或顺序前置表达轻重缓急 - 少即是多:删除重复意义的同义形容词,释放模型对核心元素的计算资源 - 稳定复现:锁定seed,微调CFG与步数,单点变量试错 现在,按以上挑战流程,将简报中的“冷与暖、静与动、湿与光”翻译为清晰可控的提示词。让蓝时的雨巷在你的指尖呼吸。
创意挑战:为新品「氮气冷萃气泡咖啡」打造主视觉海报,并进行A/B测试 挑战简介 - 你将为一款未命名品牌的「氮气冷萃气泡咖啡」创作两款风格迥异的主视觉海报,用于社媒与电商首屏A/B测试。图像应像一阵涌来的夜风,带着微凉的气泡与深烘的香气,既能在一秒内抓住目光,也能在细节处让人流连。 - 目标受众:18–35岁的城市新锐人群,偏好极简、设计感与新奇体验。 - 核心卖点:冷萃风味更纯净;氮气注入带来绵密细腻的泡沫与顺滑口感;零糖无负担。 - 构图与版式约束:请保留海报上方或右侧至少30%干净负空间用于后期排版(提示词中注明“no text”并引导留白);尽量保持产品造型统一(细长铝罐或玻璃瓶,磨砂黑/深棕主色),标签以“空白贴纸”占位,文字在后期叠加。 A/B测试假设 - 方案A(夜流体):以高速液体雕塑与霓虹反射形成强动势,强化“夜醒、酷感、城市感”的心理锚点,预期提升停留时长与分享率。 - 方案B(晨雾极简):以留白与柔光强调“纯净、顺滑、轻负担”的感受,预期提升点击转化与好感度。 A案:夜流体(动势与霓虹) - 海报文案(用于后期排版,非由SD生成) - 标题:夜色,被气泡轻轻摇醒 - 副标题:一口氮气冷萃,丝滑如夜雨落在霓虹 - 口号:顺滑,不必喧哗 - 稳定扩散提示词(正向) - 提示词(英文为主,附中文引导) hyper-detailed studio hero shot of a matte black nitro cold brew coffee can, slim cylindrical aluminum, blank label, micro-condensation beads, silky nitrogen foam crown, high-speed liquid splash sculpting around the can, deep coffee amber highlights, glossy reflections, neon rim light (blue-violet), wet floor reflections, cinematic lighting, polarizing filter look, macro texture, dramatic contrast, volumetric haze, urban night mood, rule-of-thirds composition, large clean negative space on the top-right, no text, no logo - 构图/摄影细节关键词 - camera: 85mm lens, f/4, high-speed flash freeze - lighting: key light from left, neon rim from back-right (blue-violet), glossy speculars - color palette: black, deep umber, electric blue-violet, chrome accents - 负面提示词(统一建议) - lowres, blurry, out of focus, text, letters, words, watermark, logo, distorted can, extra fingers, deformed label, bad anatomy, jpeg artifacts, oversharpen, noisy, poorly drawn, uneven foam, bent cylinder, duplicated product - 建议参数(SDXL或同等写实模型) - 模型:SDXL写实风格检查点(photoreal) - 采样器:DPM++ 2M Karras(或同等) - 步数:28–36 - CFG:5.0–6.5(推荐5.5) - 分辨率:竖版 1024x1536 或 896x1344(后期放大至海报尺寸) - 种子:固定如 321987(便于复现);若微调液体形态,可在相同种子下仅改变提示词权重 - 控制与一致性(可选) - 参考图/形状保持:ControlNet Canny 或 Depth,权重0.6–0.9;输入产品轮廓线稿,确保罐身垂直与比例准确 - 外观一致:IP-Adapter/Reference Image(色材参考)0.5–0.7,避免标签/文字被抄写,保持“blank label” - 局部修复:溢泡边缘与液体边丝可用inpaint修整,模糊的水珠进行窄范围重绘 B案:晨雾极简(留白与柔光) - 海报文案(用于后期排版,非由SD生成) - 标题:第一口,安静的气泡 - 副标题:冷萃更清透,氮气更顺滑 - 口号:清醒得很温柔 - 稳定扩散提示词(正向) - 提示词 ultra-clean minimal still-life of a matte black nitro cold brew coffee bottle with blank label, fine micro-foam at the rim, dew-like condensation, soft morning light through sheer curtain, gentle volumetric mist, pale warm highlights, wide negative space top-left, subtle tabletop reflection, premium product photography, photorealistic, crisp edges, studio-grade softness, no text - 构图/摄影细节关键词 - camera: 50mm lens, f/5.6, tripod still-life - lighting: large softbox from left, white bounce fill, backlight bloom - color palette: warm gray, matte black, soft cream, hint of coffee amber - 负面提示词 - 同A案,附加:harsh contrast, neon, busy background, clutter, heavy splash - 建议参数 - 模型/采样器/步数/CFG:与A案一致以控制变量 - 分辨率:竖版 1024x1536;为极简留白,请在提示词中强调“large negative space”与“clean background” - 种子:固定如 654123(A/B可各自固定不同种子,但同一方案内的文案微调保持同一种子) - 控制与一致性(可选) - ControlNet Depth 0.5–0.7 保持瓶型;Reference Image 0.5 统一材质与黑度 - 通过曝光/对比度的轻度变化(而非构图)生成多稿,以确保测试集中在“气质”差异 统一制作规范与交付 - 标签与文字:请在生成阶段避免任何文字,由后期在留白区域排版;提示词必须包含“no text, blank label”并在负面词中屏蔽“letters, words, logo, watermark”。 - 留白:至少30%负空间(标注“large negative space”并明确在某一侧,如“top-right”或“top-left”)。 - 色彩与质感:同一批次内黑/棕主色保持一致;冷暖对比作为A/B关键差异点。 - 导出建议:16-bit PNG(sRGB),后期统一色彩与锐化;若需大幅面,先以原图进行AI放大(2–4x)再排版。 - 命名:A_nightfluid_v01.png / B_morningminimal_v01.png,以便追踪测试版本。 可替换占位变量(便于复用到其他新品) - [容器形制]:slim can / glass bottle / stubby can - [表面材质]:matte black / dark amber glass / brushed aluminum - [液体颜色]:deep coffee amber / golden tea / clear sparkling - [情绪光]:neon rim blue-violet / golden hour warm / moonlight cool - [环境]:urban wet floor reflections / soft studio drape / foggy minimal 专业提示 - “8k, high detail”等词在提示中常作为风格暗示,并不保证真实输出分辨率;请以高分辨率生成或后期放大实现印刷需求。 - 控制变量要尽量严格:模型、分辨率、相机/光线、产品形制保持一致;A/B只变动“氛围/色温/动静”三要素。 - 若出现标签扭曲、罐身变形,优先使用ControlNet+Canny/Depth或局部重绘修正轮廓,再进行整体微调。 测试用快速清单(一次到位) - A案正向提示词(复制直用): hyper-detailed studio hero shot of a matte black nitro cold brew coffee can, blank label, nitrogen foam crown, high-speed liquid splash, coffee amber highlights, neon rim light (blue-violet), wet reflective floor, cinematic contrast, macro texture, urban night mood, rule-of-thirds, large clean negative space top-right, no text, photorealistic - A案负面提示词: lowres, blurry, text, letters, words, watermark, logo, distorted can, uneven foam, bent cylinder, artifacts, oversaturated, clutter - B案正向提示词(复制直用): ultra-clean minimal still-life of a matte black nitro cold brew coffee bottle, blank label, delicate micro-foam, soft morning backlight, sheer curtain glow, gentle mist, wide negative space top-left, crisp edges, premium product photography, photorealistic, no text - B案负面提示词: lowres, blurry, text, letters, logo, watermark, neon, harsh contrast, busy background, splash, clutter 把画面交给气味与触感:A案像在夜里捧起一团液态的雕塑,霓虹从边缘擦过,泡沫像细雪一样安静地堆叠;B案让一束清晨的光落在磨砂黑上,气泡在瓶口绽开成一圈低语。让图像先说话,再让文字轻轻落座。
Creative Challenge: “Lanterns in the Salt Fog” Premise Design a fog-swept pier-side teahouse that glows like a pocket hearth against a cobalt night. Weathered cedar glistens with ocean spray; brass lanterns bead with condensation; steam curls from porcelain like breath in winter. You can almost taste the brine, hear the gulls, and feel the damp silk of the air. This challenge asks you to architect both mood and meaning: first by distilling atmosphere into a precise mood board, then by shaping a narrative-driven concept board with a hero image and detail insets. Your language should conjure textures, light, and temperature; your images should feel like you could put your palm on the wood and leave a warm print. Deliverables - Mood Board (grid): a tactile atlas of palette, materials, light, and atmosphere. - Concept Board (hero + insets): a composed storytelling plate with a principal scene and three focused detail crops. Model + general guidance (non-binding defaults, tune to your rig) - Model: SDXL 1.0 (or a high-quality photoreal/illustration SDXL finetune). - Sampler: DPM++ 2M Karras, 30–45 steps for base; up to 60 if intricate. - CFG scale: 4.5–7.0 (lower for cohesion, higher for punch). - Aspect ratio: Mood board 16:9 or 4:3; Concept board 3:2 or 16:9 for cinematic spread. - Seed: fix seeds for reproducibility; vary per board component. - Optional: ControlNet with a simple grid sketch (Canny/SoftEdge) to lock collage; IP-Adapter Color for palette consistency; upscaler x1.5–2.0 for crisp detail. - Prompt craft: put subject and intent early; favor specific nouns over adjective soup; use weighted terms sparingly (example: wet cedar:1.2). Avoid artist names unless you have rights. Part I — Mood Board Prompt (2×4 or 3×3 grid) Goal: A clean, graphic “flat lay” showing materials, palette swatches, and light moods. No legible text—leave white space for labels you’ll add later. Prompt A: Mood Board (text-only) - Positive prompt: mood board, curated flat lay collage, neat grid layout on warm parchment paper, taped corners, clean margins, palette swatches and material studies, salt fog harbor at night mood, weathered wet cedar planks with silver lichen, hammered brass lantern lattice glowing amber, porcelain teacup with faint tea stain rings, ink-black seawater with micro ripples and oil-slick iridescence, basalt cobbles slick with rain, indigo silk ribbon damp and slightly translucent, bioluminescent blue hints in the mist, condensation droplets, soft top light, gentle shadows, high detail texture photography, cohesive color story: amber, indigo, sea-glass green, pewter gray, soot black - Negative prompt: letters, text, captions, logos, watermarks, people, faces, buildings, cluttered layout, chaotic collage, uneven borders, frames-within-frames, low-resolution, overexposed highlights, neon rainbow, plastic sheen, glossy magazine glare - Suggested parameters: Steps 40, CFG 5.5, Sampler DPM++ 2M Karras, Size 1280×800 (or 1440×960), Seed fixed (e.g., 212121) Optional tile-level micro-variations (generate as separate tiles if you prefer assembling later): - Wet cedar close-up: macro, water-beaded weathered cedar plank, silvery lichen flecks, cold salty sheen, raking light revealing grain, micro-fibers and nicks - Brass lantern filigree: macro, hammered brass lattice with warm amber glow behind, condensation droplets, soft bloom, shallow depth of field - Sea surface: top-down ink-black sea, fine concentric ripples, specks of bioluminescence, faint reflection of lantern light, low-key - Palette swatches strip: neat paint swatches: burnt amber, indigo dusk, sea-glass green, pewter fog gray, basalt black, clean paper texture, taped corners Part II — Concept Board Prompt (hero composition + insets) Goal: A narrative image that feels lived-in—warmth versus cold, hand versus ocean, lantern versus fog—plus three insets that isolate signature details. Prompt B1: Concept Board (hero image) - Positive prompt: concept board, hero composition, fog-drenched wooden pier teahouse on pylons, hand-planed cedar walls dark with rain, sliding shoji panels glowing from within, brass lanterns casting amber halos that melt into cobalt mist, steam curling from porcelain cups on a low lacquered table, wind-bent prayer ribbons in indigo silk, puddles on plank floor mirror the lantern light, gentle gull silhouettes drifting, subtle bioluminescent flecks at pier edge, low-key cinematic lighting, volumetric fog, wet specular highlights, long-lens compression of pier posts, quiet human presence: solitary figure in indigo haori seen from behind, hair stirred by sea breeze, composition with clean white margins and space reserved for detail insets, cohesive color story: amber warmth vs blue-gray cool, high fidelity, realistic textures, atmospheric perspective - Negative prompt: text, letters, signage, logos, extra people, modern city clutter, electric wires, plastic gloss, harsh flash, oversaturated neon, chaotic reflections, crooked horizon, aliased edges, smeared faces, illustrative outline halos (if going photoreal) - Suggested parameters (photoreal): Steps 45, CFG 5.0, Sampler DPM++ 2M Karras, Size 1536×960 (or 1344×768), Seed fixed (e.g., 314159), mild lens bokeh: f/2.8, 50–85mm equivalent Prompt B2: Concept Board Insets (3 detail crops, generate separately or via regional prompting) - Lantern lattice close-up: macro, brass lantern filigree, beads of condensation trembling on the lattice, inner flame diffused by fogged glass, amber glow feathered by mist, tight framing, shallow depth, tactile metal grain - Tea surface detail: macro top-down porcelain teacup, ripple of amber tea swirling in a slow gyre, faint tea oil rainbow, a single steam plume threading upward, porcelain crazing barely visible - Joinery and plank texture: low-angle macro, soaked cedar beam joinery pegged with darkened wooden dowels, droplets pooling in the seam, salt crust whispering white, raking light across rough grain Optional creative toggles - Photographic realism: add realistic, natural skin tones, physical camera language: 35mm/50mm/85mm lens, f/2–f/4, filmic contrast, soft halation - Painterly illustration: add watercolor granulation, soft edge bleed, paper tooth texture, ink wash, restrained palette, subtle brushwork, hand-drawn callout circles without text - Stylized CG: add physically based rendering, subsurface scatter on steam, ray-traced reflections in puddles, cinematic bloom, film grain 10% Control and composition tips - To lock the grid and margins: draw a simple grid (rectangles for hero + three insets) and feed it to ControlNet (Canny or SoftEdge, low-to-mid weight 0.5–0.8). Prompt “clean margins, balanced layout” for tidy hierarchy. - To keep text areas blank: include “empty labels, no legible text” in the positive prompt and “text, letters” in the negative prompt. - Palette cohesion without references: insert “limited palette: amber, indigo, sea-glass green, pewter gray, basalt black” early in the prompt. If available, use IP-Adapter Color to enforce palette. - Avoid prompt bloat: limit to concrete nouns and measurable qualities (wet cedar, condensation, raking light) rather than stacks of generic superlatives. Quality checklist (before finalizing) - Do materials read tactile at a glance? Wood grain, metal hammering, porcelain glaze, water behavior. - Is the warm–cool dialogue clear (amber vs cobalt/pewter)? - Does fog shape depth through layered contrast and soft occlusion? - Are specular highlights plausible for wet surfaces without plastic overgloss? - Do insets amplify signature design decisions (lantern, tea, joinery), not repeat the hero? Negative prompt library (curate, don’t dump) text, letters, numbers, captions, logos, watermarks, JPEG artifacts, overprocessed HDR, plastic shine, neon rainbow, messy collage, misaligned grid, lens dirt blobs, chromatic noise, AI signature smearing, duplicate subjects, disfigured hands, bland gray mush Advanced prompt skeletons (swap variables as needed) - Mood board skeleton: mood board, flat lay grid on {paper_texture}, taped corners, clean margins, palette swatches: {palette_terms}, material studies: {materials_list}, light mood: {light_mood}, atmosphere: {atmo_terms}, high detail, cohesive color story - Concept hero skeleton: concept board, hero shot of {place_asset} in {weather}, key lighting: {light_setup}, focal action: {micro_narrative}, color contrast: {warm_vs_cool}, composition: {lens_and_angle}, atmosphere: {fog_haze_terms}, texture fidelity: {materials_focus}, margins reserved for insets - Insets skeleton: macro inset of {detail_subject}, tactile surface cues: {micro_textures}, depth of field: {dof}, highlight behavior: {specular_behavior} Example variable fills - {paper_texture}: warm parchment, toothy kraft paper, muted gray museum board - {palette_terms}: burnt amber, indigo dusk, sea-glass green, pewter fog, basalt black - {materials_list}: wet cedar grain, hammered brass, porcelain glaze, basalt cobbles, silk ribbon, seawater ripples - {light_mood}: lantern amber pooling on wet wood, moonlit fog in cobalt tones - {atmo_terms}: salt spray, condensed haze, micro-droplet bloom - {place_asset}: pier teahouse on pylons - {weather}: cold marine fog with light drizzle - {light_setup}: warm interior spill vs cool ambient fog, volumetric shafts - {micro_narrative}: a solitary figure setting a cup down, steam unraveling in the draft - {lens_and_angle}: low vantage, 50–85mm, slight compression - {fog_haze_terms}: layered mist, soft occlusion, gentle falloff - {materials_focus}: realistic wet wood, fogged glass, patinated brass - {detail_subject}: lantern lattice, tea surface, beam joinery - {micro_textures}: condensation droplets, tea oil film, salt crust grains - {specular_behavior}: wet specular with soft bloom, no plastic glare - {dof}: shallow, creamy bokeh Submission suggestion - Export 1 mood board + 1 concept board (with three insets), each with consistent seed notes and parameters listed. If assembling in a layout tool, keep generated borders intact, add your labels manually for clarity. The aim is to let the image smell of sea and brass, taste of tannin and salt, and glow with a hearth’s small courage in the weather. Keep the language concrete, the palette disciplined, and the fog a sculptor of space, not a smothering veil.
将创意简报转为可生成的图像方案,迅速探索风格方向,批量产出草图与定稿,显著压缩修改与交付时间。
围绕新品卖点与人群情绪,生成主视觉、海报与详情页氛围图,多变体用于AB测试,提升点击与转化。
构建情绪板与概念板,统一叙事与调性,为拍摄或3D制作给出清晰的视觉方向与参考稿。
快速生成封面、配图与缩略图,贴合平台风格与热点语境,提高内容完读率与订阅转化。
为角色、场景与道具提供挑战式设定与气氛图,快速验证世界观与风格线,加速立项评审。
设计分层挑战题,示范从文本到图像的创作流程,帮助学生理解叙事、构图与光影表达。
将“基于挑战的场景创意设计”提示词打造为你的创意加速器:把模糊灵感转化为可落地、可复用的高质量出图指令。通过“挑战任务+场景叙事”的结构化方式,促使模型生成兼具视觉冲击、情绪张力与概念深度的Stable Diffusion提示。面向品牌营销、概念设计、插画创作与艺术教育,快速搭建清晰命题,明确风格与限制,直出多语言可用的专业级提示,并附带可操作的变体与迭代建议。目标是显著提升灵感命中率、降低试错成本、稳定风格一致性,缩短从创意到可用素材的路径。输出包含:核心命题、场景叙事、关键视觉元素、镜头与光影、情绪与风格参考、需要避免的元素、迭代与变体路径、语言设置与交付格式。
将模板生成的提示词复制粘贴到您常用的 Chat 应用(如 ChatGPT、Claude 等),即可直接对话使用,无需额外开发。适合个人快速体验和轻量使用场景。
把提示词模板转化为 API,您的程序可任意修改模板参数,通过接口直接调用,轻松实现自动化与批量处理。适合开发者集成与业务系统嵌入。
在 MCP client 中配置对应的 server 地址,让您的 AI 应用自动调用提示词模板。适合高级用户和团队协作,让提示词在不同 AI 工具间无缝衔接。
免费获取高级提示词-优惠即将到期