AI Image Prompts · 绘画提示词库
出处与授权
本站只收录项目元数据与简介,不对作者正文做转载;完整内容请到上方出处查看。商用前请自行核对原仓库的 LICENSE 与 NOTICE。
展开许可证全文
MIT License Copyright (c) 2026 YouMind-OpenLab Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
LICENSE · 2026-09-26 23:14:14
产品介绍
Recommend curated prompts from a 10,000+ real-world image generation prompt library. Works with ANY AI image model — Nano Banana Pro, Nano Banana 2, Seedream 5.0, GPT Image 1.5, Midjourney, DALL-E 3, Flux, Stable Diffusion, and more. Use this skill when users want to: - Find proven image generation prompts (any model) - Get prompt inspiration for portraits, products, social media, posters, etc. - Create illustrations for articles, videos, podcasts, or marketing content - Browse categorized prompt templates with sample images - Translate and understand prompt techniques
资源信息
许可证信息摘自官方源,可能过时或缺失。商用前请以原仓库的 LICENSE 文件为准。
安装方式
npx skills i YouMind-OpenLab/ai-image-prompts-skill
功能说明
- 按用途分成 11 类,不是随机堆砌
- 全文已入库,站内直接检索、可复制
- 语义检索(描述需求即可)
- 内容改写模式(贴正文生成定制提示词)
- 数据每日同步两次
- MIT 许可,无需 API Key
提示词库 · 2.3万
数据源:社区提示词库(MIT 许可,每日同步)共 15141 条提示词,按 11 个用途归类;同一条可归入多个用途,因此分类条数相加会大于去重后的总数。
命中 725 条
One Piece Devil Fruit Image Generation Prompt Template
game-assetA template prompt for Nano Banana Pro/2 that allows users to generate images of themselves having eaten a specific Devil Fruit from One Piece, such as the Mera Mera no Mi (Flame-Flame Fruit). Users simply replace the placeholder with the name of the desired Devil Fruit.
展开提示词全文
It's okay to insert the name of your favorite character's Devil Fruit into the first <<DEVIL FRUIT NAME>> in the prompt from the source.
Example: <<{argument name="Devil Fruit Name" default="Mera Mera no Mi"}>>
One Piece Devil Fruit Transformation Template
game-assetA complex, structured prompt template designed for Nano Banana Pro to transform a photo of a real person into a photorealistic One Piece Devil Fruit ability user. The prompt requires the user to input a Devil Fruit name and details how the AI should interpret and apply the ability (e.g., flames, ice, stretched limbs) in a dynamic, live-action style while strictly preserving the subject's face and background.
展开提示词全文
{
"version": "1.0",
"description": "Transform a real person in a photo into a One Piece-style Devil Fruit ability user: dynamic pose, dramatic camera angle, ability-specific SFX, and a photorealistic manifestation of the specified Devil Fruit’s power.",
"variables": {
"DEVIL_FRUIT_NAME": {
"type": "string",
"value": "<<DEVIL FRUIT NAME>>",
"description": "The canonical Devil Fruit name whose ability should be interpreted and applied to the subject (e.g., Gum-Gum Fruit, Flame-Flame Fruit, Ice-Ice Fruit, Sand-Sand Fruit, Rumble-Rumble Fruit, Flower-Flower Fruit, Dark-Dark Fruit, Paw-Paw Fruit, etc.)."
}
},
"steps": [
{
"name": "DEVIL_FRUIT_ABILITY_EDIT",
"goal": "Apply a dynamic One Piece-style pose, dramatic camera angle, ability-specific manga SFX, and a photorealistic manifestation of the specified Devil Fruit’s ability.",
"instructions": "Use the uploaded photo as the base.\n\nKeep the person photorealistic (eyes, face, skin texture, hair, makeup). Do NOT convert the scene into manga. Do NOT replace the background.\n\n=====================================\nDEVIL FRUIT ABILITY TRANSFORMATION\n=====================================\nInsert your Devil Fruit name here: {argument name="Devil Fruit Name" default="[DEVIL_FRUIT_NAME]"}\n\nInterpret the canon ability of [DEVIL_FRUIT_NAME] realistically and apply it to the body or surrounding environment. The effect must look like a live-action One Piece movie.\n\nExamples:\n- Gum-Gum Fruit: stretched limbs, elastic skin, tension folds, rubber reflection\n- Flame-Flame Fruit: realistic flames, heat distortion, warm illumination\n- Ice-Ice Fruit: frost buildup, ice crystals, cold vapor\n- Rumble-Rumble Fruit: lightning arcs, photoreal sparks\n- Sand-Sand Fruit: sand disintegration, sandstorm motion\n- Flower-Flower Fruit: photoreal arms blooming from surfaces\n- Paw-Paw Fruit: shockwave compression ripples\n- Dark-Dark Fruit: gravitational distortion, dark smoke tendrils\n\nRules:\n- The face must remain recognizable.\n- Effects must integrate into real lighting and shadows.\n- No cartoon filters — supernatural physics must look real.\n\n=====================================\nONE PIECE-STYLE POSE TRANSFORMATION\n=====================================\nRepose the body into a dramatic battle pose:\n- wide stance, strong dynamic gestures\n- stretched arm for Gum-Gum, open palm for shockwave, ignition stance for flame, elemental charge-up for logias, etc.\n- clothing should deform."
Enhancing Three-View Drawings and Generating Angles with Nano Banana
game-assetThe user describes a two-step process: first, using Nano Banana Pro to enhance a three-view drawing (specifically interpreting a mysterious neck decoration as a 'magic-infused stone'), and second, using Novel AI to generate images with various angles and expressions based on the enhanced design for training material.
展开提示词全文
First, use the power of Nano Banana to enhance the three-view drawing. Nano Banana seems to have interpreted the mysterious decoration around the neck as a 'magic-infused stone.' Since that kind of decoration is acceptable, I decided to proceed with it (the decoration in this area varies depending on my mood, so I will fix it as a magic-infused stone from now on). Then, use Novel AI's features to generate images with various angles and expressions. Use these to prepare the minimum necessary learning materials.
X-Ray/Thermal Visualization Prompt for Mecha (VOTOMS Scopedog)
game-assetA highly detailed prompt for Nano Banana Pro designed to generate a technical, high-contrast visualization of a mecha (Armored Trooper Scopedog from VOTOMS) using a hybrid X-ray and thermal heatmap aesthetic. It specifies orthographic projection, internal component details (reactor, hydraulics), and realistic battle-worn surface textures, emphasizing mechanical realism and technical manual style.
展开提示词全文
subject: "Armored Trooper ATM-09-ST Scopedog, classic first-generation AT from Armored Trooper VOTOMS"
view: "tight frontal view, perfect orthographic projection, zero perspective distortion, strict symmetrical composition"
style: "high-contrast military-industrial scientific render, cold war-era weapons lab visualization"
background: "pure black void, absolute vacuum environment, no reflections, no floor, no shadows except self-emission"
aesthetic: "false-color thermal X-ray hybrid imaging with heavy mechanical wear"
transparency:
level: "semi-transparent heavy armor plating and rolled homogeneous steel composite"
reveal: "extremely sharp and brutally detailed internal skeleton and systems"
internal_components:
- "central torso-mounted turbo-custom plasma fusion reactor"
- "exposed hydraulic muscle cylinders and heavy-duty linear actuators"
- "high-pressure fluid lines pulsing with red-hot coolant"
- "belt-fed 30mm gatling arm punch with visible ammo feed"
- "rear skirt-mounted high-thrust rocket thrusters"
- "dense array of infrared/optical sensors and lens turret"
- "cockpit block with visible pilot seat and control sticks"
- "knuckle joints with exposed wiring and coolant pipes"
heatmap_gradient:
base: "deep cold blues and gunmetal grays (inactive armor zones)"
mid: "dull green to burnt orange (hydraulic pressure zones)"
high: "glowing yellow to molten red (reactor core, thruster bells, gatling barrel)"
peak: "blinding white-hot hotspots on reactor shielding and over-pressurized joints"
lighting:
primary: "harsh cyan-white plasma glow from reactor vents"
sources:
- "chest reactor glare bleeding through armor seams"
- "red-hot hydraulic lines"
- "amber warning strobes on lens turret"
- "molten orange glow from thruster nozzles"
secondary: "faint rim lighting on battle-scarred armor edges, subtle volumetric heat haze"
surface_details:
- "authentic battle damage, bullet dents, scorch marks, chipped paint"
- "visible weld lines, rivets, maintenance hatches"
- "grimy industrial texture, oil stains, hydraulic fluid leaks"
quality_boosters:
- "8K resolution, ultra-detailed mechanical realism"
- "hyper-sharp engineering precision, Yamato 1/12 model kit level detail"
- "military technical manual aesthetic, cold and unforgiving"
- "photorealistic metal shading, micro-scratches visible"
negative_prompt: "cartoon, anime exaggeration, clean toy look, glowing eyes, pilot face visible, green android, Gundam parts, bit staves, permet effects, logo, text, watermark, extra limbs, perspective distortion, bright background, cute, deformed"
weight_guidance:
"(masterpiece, ultra-detailed, absurdres, best quality):1.4"
"(Armored Trooper Scopedog, perfect proportions, authentic VOTOMS design):1.6"
"(military industrial X-ray thermal aesthetic, brutal mechanical realism):1.35"
"(battle-worn h
Ultimate Weapon Mastery Cinematic Pose Prompt
game-assetA complex JSON-structured prompt for Nano Banana Pro designed to automatically analyze an uploaded weapon's type (e.g., light blade, heavy blade, firearm) and generate a highly cinematic, heroic action pose. It incorporates dynamic foreshortening, low-angle hero shots, and contrast-heavy lighting to maximize lethality and aesthetic impact, ensuring the resulting image looks like a key visual from a movie or game.
展开提示词全文
{
"preset_name": "Ultimate_Weapon_Mastery",
"mission": "MAXIMIZE_LETHALITY_AND_AESTHETICS",
"source_integrity": {
"_hint": "[PREMISE] Maintain the character's identity, but actively allow for dramatic changes like 'sharp gaze' or 'wind-blown hair' typical of combat situations.",
"identity_lock": "strict_facial_features",
"combat_mode_adaptations": {
"expression": "intense_focus_or_fearless_smile",
"hair_physics": "dynamic_flow_matching_action_speed",
"clothing_damage": "optional_minor_wear_and_tear_for_realism"
}
},
"weapon_analysis_engine": {
"_hint": "[WEAPON ANALYSIS] Analyze the 'type', 'center of gravity', and 'method of killing' of the weapon provided to determine the optimal stance.",
"analyze_weapon_class": true,
"tactical_mapping_logic": {
"class_blade_light": {
"keywords": ["katana", "dagger", "saber"],
"action_style": "high-speed_slash_or_iaijutsu_stance",
"camera_effect": "motion_blur_on_blade"
},
"class_blade_heavy": {
"keywords": ["greatsword", "axe", "hammer"],
"action_style": "weighty_impact_pose_using_ground_drag_or_shoulder_rest",
"camera_effect": "camera_shake_simulation"
},
"class_firearm_handgun": {
"keywords": ["pistol", "revolver", "dual_wield"],
"action_style": "close_quarters_combat_stance_(C.A.R._system)_or_dynamic_aim",
"grip_discipline": "trigger_discipline_active"
},
"class_firearm_long": {
"keywords": ["rifle", "sniper", "shotgun"],
"action_style": "stable_aiming_with_cheek_weld_or_patrol_carry",
"perspective": "looking_down_the_sights"
},
"class_polearm": {
"keywords": ["spear", "halberd", "staff"],
"action_style": "dynamic_thrust_with_extended_reach",
"composition_note": "utilize_diagonal_lines_across_frame"
}
}
},
"heroic_pose_database": {
"_hint": "[CHOREOGRAPHY GENERATION] Refer to 'key visual' level pose data from movies, games, and anime to eliminate amateurish results.",
"retrieval_logic": {
"source": "cinematic_combat_pose_db",
"query_parameters": [
"weapon_type",
"character_physique",
"desired_coolness_factor"
],
"pose_correction": {
"line_of_action": "strong_s_curve_or_c_curve_body_line",
"silhouette_check": "ensure_weapon_is_clearly_visible_against_background"
}
},
"dynamic_foreshortening": {
"_hint": "[FALSE PERSPECTIVE] Exaggerate the size of weapons or limbs closer to the camera to enhance impact (perspective emphasis).",
"apply_perspective_distortion": "moderate_to_high",
"focus_point": "weapon_tip_or_muzzle"
}
},
"cinematic_combat_composition": {
"_hint": "[DIRECTION] Create the 'atmosphere of a battlefield', not just a standing figure.",
"camera_work": {
"angle": "dutch_angle_to_create_tension",
"height": "low_angle_hero_shot",
"lens_choice": "wide_angle_24mm_for_depth_exaggeration"
},
"visual_effects": {
"lighting": "contrast_heavy_rim_lighting",
"atmosphere": "sparks_dust_or_energy_aura_matching_weapon",
"depth_cue": "blurred_foreground_elements_(debris/sparks)"
}
},
"safety_a
Dynamic Wide-Angle Action Pose Prompt (Gatotsu Style)
game-assetA highly detailed, structured prompt for generating a dynamic, anime-style illustration of a character in a specific action pose, emphasizing extreme perspective distortion (super wide-angle lens) and dramatic camera angles (Dutch angle). The prompt meticulously defines the scene, character appearance, outfit, pose, and the exaggerated perspective of the hands, aiming for a high-impact visual effect.
展开提示词全文
Scene:
type: "battle scene"
camera_angle: "dramatic dutch angle, tilted composition"
lens:
type: "super wide-angle lens"
effects:
- "extreme perspective"
- "strong spatial compression between foreground, midground, background"
Character:
identity:
ethnicity: "Japanese"
gender: "{argument name="gender" default="girl"}"
hair: "{argument name="hairstyle" default="Brown medium length hair with cat ears"}"
outfit:
style: "windy, dressy, Idol costumes that resemble high school girl uniforms"
pose:
description: "bold, dynamic action pose"
body_orientation: "leaning with momentum implied by the tilt"
Hands_and_Gesture:
right_hand:
gesture: "holding a Japanese sword"
position: "held in front of her face, higher than her left hand, blade down and pointing towards you"
gaze_alignment: "her eyes follow the pointing direction"
left_hand:
gesture: "open hand"
fingers: "all five fingers widely spread"
position: "The palms are facing forward as if aiming."
palm_direction: "facing upward as if supporting it"
perspective_role: "dominates the extreme foreground"
perspective:
dominance: "left hand is massively enlarged by wide-angle distortion"
clarity: "both hands and face in sharp detail"
depth_of_field: "full DOF, no background blur"
Lighting:
type: "soft indoor lighting"
mood: "gentle highlights, controlled shadows"
background: "neutral, blurred environment"
Cinematic_Effects:
atmosphere: "action tension + stylized anime dynamism"
motion_sense: "perspective and Dutch angle add momentum"
Output:
aspect_ratio: "16:9"
style: "high-quality anime illustration"
rendering: "clean lines, sharp details, accurate anatomy"
Persona-Style Status Screen Generation
game-assetA prompt used to generate a status screen in the style of the Persona video game series, instructing the AI to randomly generate high values for the level and stats.
展开提示词全文
A Persona-style status screen using nano banana pro. I instructed it to randomly generate the level and stats, but they tend to be high.
3D Polished Character Portrait (Pastel Anime Style)
game-assetA Nano Banana Pro prompt for generating a highly polished 3D character portrait in a soft pastel, toy-like aesthetic, emphasizing smooth surfaces, gentle lighting, and an expressive, anime-inspired facial structure.
展开提示词全文
Create a highly polished 3D character portrait in a soft pastel, toy-like style. Smooth surfaces, gentle lighting, and clean gradients. The avatar should have expressive eyes, stylized hair, soft shadows, and a friendly, anime-inspired facial structure. Rendered in high-quality CGI with subtle ambient glow.
Isometric 3D Tony Stark Workshop Cube
game-assetA complex, highly detailed prompt for generating an isometric 3D cutaway cube miniature room depicting Tony Stark's underground workshop during the autonomous assembly of the Iron Man Mark 42 suit, emphasizing chibi/figurine style, dramatic lighting, and photorealistic materials.
展开提示词全文
An isometric 3D cube-shaped miniature room (perfect shallow cutaway cube, everything 100% strictly contained inside the cube boundary, no protrusion). Room description: Tony Stark's underground workshop during the Iron Man Mark 42 suit autonomous assembly process. Dimly lit high-tech garage lab: brushed steel and matte black walls covered with holographic blueprints, arc reactor glow, scattered tools (DUM-E robotic arm holding a coffee mug, U robotic arm spraying fire extinguisher), multiple half-assembled red-gold Iron Man suits hanging on robotic arms, a central rotating platform where the Mark 42 suit pieces are flying and auto-assembling onto a dummy Tony, sparks flying, welding arms active, floor littered with micro thrusters, faceplates, repulsor parts, empty whiskey glasses, and a glowing circular arc reactor prototype on the workbench. Wall has a large "Stark Industries" logo and Jarvis interface screens showing suit diagnostics. Character: chibi/figurine-style Tony Stark in black tank top, perfectly groomed goatee, matte PVC material, big head small body proportion, standing on the workbench with arms outstretched, confidently directing the autonomous suit pieces to assemble around the dummy, cocky satisfied smirk, one eyebrow raised, wearing welding goggles pushed up on forehead. Lighting: dramatic cinematic workshop atmosphere (cold blue holographic glow from screens + intense white welding sparks + signature arc reactor ice-blue light + warm orange workshop lamps), realistic metallic reflections, colored caustics and god rays cutting through light smoke. Camera: slightly elevated isometric three-quarter view, front edge of the cube perfectly centered and framed, shallow depth of field, foreground sharp, background softly blurred, photorealistic materials, ultra-high detail on every screw and reflection, clean museum-like composition, neutral dark backdrop, no text, no watermark, no elements outside the cube. 8K, masterpiece, best quality, Gemini 3 Pro Nano Banana Pro style
Pokémon Style Fusion Promo
game-assetA prompt for generating a style-fused Pokémon image combining retro GameBoy pixel art, 3D models, and watercolor landscapes in a single frame.
展开提示词全文
With Pokémon StyleFusion, it's easy to mix retro GameBoy pixel art, 3D Pokémon models, and watercolor landscapes. In this shot, pixel-{argument name="first pokemon" default="Charmander"}, 3D-{argument name="second pokemon" default="Bulbasaur"}, and a watercolor {argument name="third pokemon" default="Pikachu"} appear in one frame.
Collectible Art Toy Character Design Sheet
game-assetA prompt for generating a design sheet for a collectible art toy character. It specifies the character's appearance (fluffy, cute, round face, yellow safety helmet, blue glasses), the required views (front, side, back), the style (super cute 3D, plushie figure), and technical details (pastel color scheme, clean beige background, soft studio lighting).
展开提示词全文
A collectible art toy character design sheet, front, side, and back views, fluffy , cute, round face with [{argument name="head accessory" default="a yellow safety helmet"} on his head, {argument name="eye accessory" default="blue round glasses"}], super cute 3D style, soft and tactile like plushie figure, pastel color scheme ({argument name="color scheme" default="black, cream, light yellow"}), clean solid beige background, soft studio lighting, high detail
Image Analysis and Correction Tool Description
game-assetThis tweet describes a feature of Nano Banana Pro: a game or tool that automatically analyzes and corrects images, suggesting that the prompt itself is a meta-instruction for the AI's function rather than a visual generation prompt.
展开提示词全文
A game that automatically analyzes and corrects images.
提示词模型无关,可直接用于 Midjourney / DALL·E / Flux / Stable Diffusion 等任意文生图模型。
依赖说明
- Node.js 18+ - 无需任何 API Key(数据全部公开) - 提示词已完整同步到本地与数据库,站内检索不依赖第三方
适用平台
更新记录
官方源监视中内容有更新(细节请以官方文档为准)
内容有更新(细节请以官方文档为准)
内容有更新(细节请以官方文档为准)
官方源有内容更新(首次纳入监控)
信息边界说明
本页信息摘自官方文档,仅作技术信息呈现,不代表本站的判断或推荐。价格、配额、可用性以官方最新说明为准。你离开本站后的任何行为(申请 Key、付费、部署、商用),本站不做任何承诺与担保。
同类推荐
视频短视频图片音乐音频配音口播数字人全能创作
官方标注收费通过同一个AI创作工具完成AI图片、视频、音乐与语音创作,查找公开社交数据,编辑视觉成果,并集中查看和管理生成资产。Beatra 全能创作支持AI图片生成(文生图)、AI视频生成(图生视频)和AI视频编辑,通过AI音乐生成把想法做成歌曲或纯音乐,也可用AI语音生成(文字转语音)、多语言旁白和AI语音克隆完成跨媒介内容;还能查找抖音数据、小红书数据,以及 YouTube、TikTok、Instagram 与 X 上的公开帖子、笔记、评论、账号和趋势,把社交媒体数据用于竞品调研和创作。它是一款面向AI内容创作与多模态内容创作的AI创作套件,用起来就是一个把图片、视频、音乐与语音收在同一处的AI创作助手,适用于社交内容、产品展示、营销素材、课程、播客、短片和跨媒介项目;还能在不同媒介间复用原始素材、已有成果和查到的公开资料,查看制作进度、返回文件、实际用量和积分记录。
AI数字人口播与虚拟主播视频
官方标注收费用一张人像和一段短脚本或语音,制作有明确讲述内容的数字人口播视频。可使用现成录音,也可以先选择音色,把脚本生成旁白,让画面中的人物随声音完成稳定、自然的口播。适合 AI 主播、虚拟主播视频、数字人讲解、让照片说话、照片说话视频、产品讲解视频、数字人培训视频、公告和社交口播,并在交付后检查人物一致性、声音清晰度、口型同步和动作稳定性。一节课的讲解也能做成课程口播。
罗斯蒙德可视化图形生成
官方未说明【罗斯蒙德可视化图形生成】南京罗斯蒙德基于已交付项目沉淀的可视化图形生成场景:根据业务描述或系统需求,自动生成可编辑的架构图、流程图、ER 图、时序图、状态图与 C4 图。采用「统一图模型」一次建模、多格式输出:结构化文本(Mermaid / PlantUML,可嵌入 Markdown、进 Git 版本管理)、可编辑图形文件(drawio / Excalidraw,可拖拽修改节点与样式)、静态图(SVG 自渲染,PNG 给转换命令),并输出图形说明(图的含义 + 关键节点解释)。内置样式系统统一配色与命名,内置布局引擎自动排布,校验脚本拦截孤立节点、悬挂连线与语义缺漏。触发词:画架构图、画流程图、画 ER 图、画时序图、画状态图、C4 图、系统架构图、部署架构图、数据流程图、实体关系图、生成 drawio、生成 Excalidraw、导出 mermaid、导出 PlantUML、导出 SVG、导出 PNG、可编辑的图、图转可编辑、架构图可编辑、流程图二次修改、画图、出图、画个图。
海报设计工作室
免费增值需申请 Key将一段活动描述、一张产品照片或品牌参考图,转化为抓人眼球的活动海报、宣传 banner 或社交媒体配图。这款 AI海报设计与制作工具生成图片丰富的海报,搭配清晰视觉层级、按品类匹配的风格和预留文字区域,覆盖活动、音乐节、电影、产品发布和促销场景。你可以把它当作 AI宣传单制作工具来制作促销宣传单,当作活动海报设计工具,当作电影海报生成器,当作音乐节海报设计工具,当作产品发布会海报的 banner设计 AI,或当作平面设计海报工作室。从活动主题出发,用品牌风格参考图变换产品或场景照片,或在已认可草稿上做定向精修,完成可印刷或可分享的成品。
架构图一键生成
官方标注收费需申请 Key使用 ContextWeave 生成或修改架构图、流程图、思维导图和复杂信息图。适用于需要从代码、文件或自然语言中提取结构与关系并实际产出 CW、SVG、HTML、可编辑 PPTX 或原生 VSDX 的请求;未特别说明时,PPTX 与 Visio 使用原生形状和连接器导出。不适用于统计图表、手绘插画或像素级排版。
自动化克隆数字人口播混剪视频
免费增值需申请 Key抖音短视频双路径生产体系(单 Skill 版)。Path A[付费·高质量]:策略大脑→(可选:素材混剪/AI高光提取)→脚本优化→AI生图/素材准备→配音渲染(百炼/GPT-SoVITS/CosyVoice/VoxCPM2+剪映/Remotion/ffmpeg/数字人)。Path B[免费·动画风]:采集热点→脚本→edge-tts免费配音+HyperFrames开源渲染→MP4,零云费开箱快。触发词分层:①"抖音短视频/选题规划/合规闸门/运营策略" ②"口播脚本/AI生图/AI配音/视频渲染/Remotion渲染" ③"蹭热点/快速出片/免费做视频" ④"混剪/高光提取/素材粗剪/autoclip
什么是技能
技能(Skill)是封装好的单项能力,比如文字识别、文档摘要、关键词提取。开发者把调用逻辑和提示词打包,使用者像用插件一样加载即可,无需了解底层模型。启枢技能 收录大量技能,可直接试跑或编入方案。