style-fusion-prompter

Style Fusion Prompter (cross-media style fusion prompt generator)

Focuses on generating high-tension, cross-media dual-style fusion prompts with coexisting **"character visual language × scene visual language"**. Firmly opposes mediocre compromise, ensuring that the two sets of visual languages for character and scene are independently distinct while being highly unified in physical space, lighting and shadows, contact, and narrative.

---

Core Design Philosophy

1. Dual languages coexist independently; refuse to flatten into mediocrity：

   • The image must simultaneously contain two sets of clearly distinguishable visual media textures;
   • The character part faithfully follows the form, lines, materials, and media feel of the [character visual language];
   • The scene part faithfully follows the perspective, lighting and shadows, texture, and environmental feel of the [scene visual language];
   • It is strictly forbidden to average the two into an ordinary mixed-style illustration, and strictly forbidden for them to fully assimilate each other.

2. Miniature photography perspective and scale contrast (core instruction)：

   • The instruction "miniature photography, character size abnormally small, scene remains at normal size" must be injected, and placed in a key position relatively early in the prompt;
   • This creates a highly dramatic miniature scale and spatial depth of field: the character lives like a miniature entity in a real or illustrated world at normal scale, greatly enhancing the media contrast and visual interest.

1. Unified physical logic; refuse sticker collage：

   • The two visual languages maintain a sharp contrast, but sharethe same physical world: unified light source direction, color temperature, weather, environmental reflections, and atmospheric perspective;
   • Between the character and the scene, there must be credible physical contact, standing center of gravity, force transmission, front-back occlusion, and cast shadows;
   • It is by no means "pasting a character texture onto a background of another style"; rather, the character truly lives within this scene.

2. Clear narrative moment, rigorous action logic：

   • The image must form a clear narrative moment around the [theme], with priority given to making "who, where, and what they are doing" immediately clear;
   • If there are actions such as running, jumping, pulling, pushing, or climbing, the support of the center of gravity and the direction of force application must conform to physical common sense; action logic takes precedence over mere exaggeration.

3. Principles of theme and style fidelity (precisely routed by input type)：

   • Theme fidelity: The theme input by the user is the core soul of the image, and must be strictly kept restrained and faithful,it is strictly forbidden to arbitrarily expand complex plot actions, character appearance, lengthy storylines, or background stories. If the user provides a short theme (such as "post-apocalyptic superpowered individual"), the field strictly uses "post-apocalyptic superpowered individual" or the user's original intended phrase;
   • Style name input (follow the user's wording; strictly forbidden to expand): If the user writes astyle name(such as "realistic", "realistic style", "cinematic live-action photography", "ink gongbi", "retro American comic", or any other name),output exactly as the user wrote it(For example, if the user specifies "realistic", output only "realistic"),Adding embellishments without authorization is strictly prohibited; expanding any specific environmental scenes, depth of field, lighting and shadow, or camera descriptions is strictly prohibited；
   • Style number input (copy the image-generation features completely and verbatim; compression is strictly prohibited): If the user has filled instyle number(such as #018、#239、240 , etc., or a number selected through AI intelligent recommendation),it must be fully expanded strictly according to the standard image-generation method in the library： #{number} {generation_name}。参考作者/风格名称：{reference}。核心风格特征：{traits}。 [Absolute red line] Core style features {traits} must be directly and completely copied word for word from styles.json ; manual compression, distillation, summarization, deletion, or rewriting is absolutely strictly prohibited! (The English corresponding complete copy is:#{number} {generation_name}. Reference author/style: {reference}. Core style traits: {traits_en}.）。

4. Emotion standardization, by default sourced from the splicer preset library：

   • Unless the user explicitly specifies an emotion word in the requirements, the [Emotion] fieldmust and can only be selected from the 15 standard presets built into the [Prompt Splicer] in the entire library： 治愈 (Healing)、童趣 (Childlike)、松弛 (Relaxed)、幽默 (Humorous)、诗意 (Poetic)、浪漫 (Romantic)、活力 (Vibrant)、微丧 (Melancholy)、孤寂 (Solitary)、紧张 (Tense)、庄严 (Solemn)、荒诞 (Absurd)、恐怖 (Eerie)、神秘 (Mysterious)、激烈 (Intense)；
   • It is strictly prohibited to invent non-standard compound long-phrase words on your own (such as "oppressive, bleak, do-or-die狂暴 epic feeling and survival awakening").

---

Input Specifications and Style Parsing

0. Activation Method and Trigger Words

• Instruction activation: The user sends an instruction beginning with 【风格融合设计】 or 风格融合设计 , which directly activates this skill;
• Prompt Assembler Standard Format： 【风格融合设计】，角色风格：279，场景风格：写实，主题色：C-01，画幅比例：3:4，情绪：治愈，主题：山中的妖精 or with a default selection:【风格融合设计】，角色风格：279，场景风格：写实，其他你帮我选择，主题：山中的妖精。

1. Input Dimensions

The user may provide the following information (free combinations are supported):

• Character style: style name (according to the user) or full-library style number (#001–#331, expanded according to the image-generation method);
• Scene style: style name (according to the user, such as "realistic", "cinematic live-action photography") or full-library style number (#001–#331, expanded according to the image-generation method);
• Theme: the specific theme of the image, which must be faithfully preserved during generation; arbitrarily expanding or fabricating plot actions is strictly prohibited;
• Emotion / Atmosphere: by default, it must be selected from the 15 standard presets of the Prompt Splicer (healing / childlike / relaxed / humorous / poetic / romantic / energetic / slightly melancholic / lonely / tense / solemn / absurd / horror / mysterious / intense), unless the user explicitly specifies otherwise;
• (Optional) Image type: 165 image-type layout numbers (SC-001–SC-023, IG-001–IG-035, SB-001–SB-071, IP-001–IP-013, EC-001–EC-024）；
• (Optional) Theme color: 36 classic theme color numbers (C-01–C-36）；
• (Optional) Aspect ratio: default 3:4, or specified as needed (such as 16:9、9:16、1:1、5:2、2.35:1）。

2. Intelligent Recommendation Strategy (When Not Specified)

• If the user provides only the theme: the AI combines the theme semantics and, from the full library of 331 hand-drawn styles and realistic styles,actively recommends a [character style × scene style] pairing with the most contrasting beauty and tension(the recommended numbers are expanded according to the image-generation method), selects the most matching emotion from the 15 presets, recommends 1 theme color, gives concrete recommendation reasons, and then directly outputs the complete prompt.
• If the user specifies only the character style: Preserve the character style (if it is a name, follow the user's; if it is a number, expand it according to the image generation method). AI combines the theme to recommend the scene style with the most dramatic contrast.
• If the user specifies only the scene style: Preserve the scene style (if it is a name, follow the user's; if it is a number, expand it according to the image generation method). AI combines the theme to recommend the most fitting character style.

---

[!IMPORTANT] [The core coexistence specification paragraph must be copied completely verbatim] In the prompt, the entire coexistence control instruction from "macro photography, the character is abnormally small in size, while the scene maintains normal size. The image must simultaneously contain two sets of clearly distinguishable visual languages" to "the themed event is clearly readable" is the underlying core contract that enables image generation models (such as GPT Image, Midjourney, Flux, etc.) to accurately understand the decoupled coexistence of dual visual media and miniature perspective. Every time a style fusion prompt is output,it must be copied exactly, word for word, in full, and attached to the latter part of the prompt. Any form of deletion, condensation, compression, tampering, or omission is strictly prohibited!

Chinese standard prompt

生成一幅“角色视觉语言 × 场景视觉语言”共存的跨媒介融合画面。缩微摄影，角色尺寸异常小，场景保持正常尺寸。
- 【角色视觉语言】：{若写名称则按用户的来，若填编号则按生图方式展开：#{number} {generation_name}。参考作者/风格名称：{reference}。核心风格特征：{traits}。}
- 【场景视觉语言】：{若写名称则按用户的来，如：写实；若填编号则按生图方式展开：#{number} {generation_name}。参考作者/风格名称：{reference}。核心风格特征：{traits}。}
- 【主题】：{用户输入的主题，忠实保留，严禁随意扩写大段剧情动作}
- 【情绪】：{默认从15种标准情绪预设库中选取，例如：激烈 / 孤寂 / 紧张；除非用户显式指定}
[- 【图型】：{可选图型编号与排版特征，若无则省略本行}]
[- 【主题色】：{可选主题色编号与名称，若无则省略本行}]
[- 【画幅比例】：{画幅比例，默认 3:4}]

缩微摄影，角色尺寸异常小，场景保持正常尺寸。
画面中必须同时存在两套清晰可辨的视觉语言。
不要把两者平均磨成普通的混合风格插画。
角色部分使用【角色视觉语言】表现，场景部分使用【场景视觉语言】表现。这里的“场景”包括环境空间、地形、建筑、植物、天空、水面、天气、地面、道具以及整体空间氛围。
【角色视觉语言】与【场景视觉语言】都必须忠实保留各自的核心特征，包括但不限于：造型逻辑、比例系统、线条方式、笔触特征、体块组织、几何倾向、材质表达、表面肌理、色彩体系、明暗方式、细节密度、空间处理方式、平面化或立体化程度以及各自独有的媒介感。不要额外强行加入与原风格无关的统一化修饰。
两种视觉语言必须保持明显差异，但共享同一个空间、光源、色温、天气、空气、构图和叙事时刻。
视觉统一应通过遮挡关系、接触关系、投影关系、地面关系、前后空间关系、局部反光、环境综合色和空气透视来完成，而不是把两种视觉语言磨成同一种质感。
角色必须真正存在于场景中，与场景形成自然互动，不能像贴纸一样浮在画面上。角色与场景之间需要有清楚可信的接触、站立、受光、投影、遮挡和空间关系。重点不是“一个角色站在另一个风格的背景前”，而是让角色真正进入并生活在这个场景世界中。
画面必须围绕【主题】形成一个明确的叙事瞬间，优先保证“谁、在哪里、正在做什么”清楚可读。不要为了展示风格而加入大量与主题无关的装饰元素。
如果画面中存在明显动作，如跑、跳、扑、拉、推、攀爬、追逐、搏斗、搬运或其他动态行为，需要保证发力点、重心、支撑关系、接触位置、受力方向、物体运动方向、遮挡和透视合理，动作逻辑优先于单纯夸张效果。
不要生硬拼贴，不要左右分栏，不要上下分区，不要贴纸叠加，不要主体悬浮，不要错误遮挡，不要不同光源互相冲突，不要让角色风格被完全同化成场景风格，也不要让场景风格被完全同化成角色风格。
最终效果应呈现：
- 缩微摄影比例，角色尺寸微小而场景保持正常尺寸
- 两种不同视觉语言自然存在于同一个世界中
- 角色风格与场景风格差异清晰
- 空间与光线逻辑统一
- 角色与场景互动自然
- 主题事件明确可读

English Standard Prompt

Generate a cross-media fusion artwork where "Character Visual Language × Scene Visual Language" coexist. Miniature photography, the character scale is exceptionally small, while the scene maintains normal scale.
- [Character Visual Language]: {If style name is provided, use user's exact name; if style number is provided, expand in generation format: #{number} {generation_name}. Reference author/style: {reference}. Core style traits: {traits_en}.}
- [Scene Visual Language]: {If style name is provided, use user's exact name (e.g., Realistic photography / Cinematic photorealism); if style number is provided, expand in generation format: #{number} {generation_name}. Reference author/style: {reference}. Core style traits: {traits_en}.}
- [Theme]: {User's theme, strictly preserved without arbitrary expansion}
- [Mood]: {Default selected from the 15 standard presets: Healing / Childlike / Relaxed / Humorous / Poetic / Romantic / Vibrant / Melancholy / Solitary / Tense / Solemn / Absurd / Eerie / Mysterious / Intense, unless specified by user}
[- [Layout]: {Layout ID and description, omit if none}]
[- [Theme Color]: {Theme color ID and name, omit if none}]
[- [Aspect Ratio]: {Aspect ratio, default 3:4}]

Miniature photography, the character scale is exceptionally small, while the scene maintains normal scale.
Two clearly distinguishable visual languages must coexist in the image simultaneously.
Do NOT average or blend the two into a generic hybrid illustration.
The character elements must be rendered strictly in [Character Visual Language], while the scene elements must be rendered strictly in [Scene Visual Language]. Here, "scene" includes environmental space, terrain, architecture, vegetation, sky, water, weather, ground, props, and overall spatial ambiance.
Both [Character Visual Language] and [Scene Visual Language] must faithfully retain their respective core characteristics, including but not limited to: modeling logic, proportional systems, linework methods, brushstroke traits, volume organization, geometric tendencies, material expressions, surface textures, color systems, shading/lighting techniques, detail density, spatial treatment, degree of flatness vs. three-dimensionality, and their distinct tactile media sensations. Do NOT artificially force any uniform stylistic modifications irrelevant to each original style.
The two visual languages must maintain noticeable contrast, yet share the exact same space, light source, color temperature, weather, atmosphere, composition, and narrative moment.
Visual unity must be achieved through occlusions, physical contact, contact shadows, ground contact, fore/background depth, subtle bounce light, environmental ambient color, and aerial perspective, rather than blending the two visual languages into an identical material texture.
The character must truly exist and naturally interact within the scene world, rather than floating like a detached sticker. There must be credible physical contact, grounded posture, lighting, cast shadows, occlusions, and depth relationships between the character and the environment. The focus is NOT "a character simply standing in front of a different-styled background", but rather having the character truly inhabit and live within this environment.
The image must center around [Theme] to form a distinct narrative moment, prioritizing clarity of "who, where, and what they are doing". Do NOT introduce clutter or irrelevant decorative elements merely to exhibit the styles.
If dynamic actions are present (e.g., running, jumping, leaping, pulling, pushing, climbing, chasing, wrestling, carrying, or dynamic movements), ensure the center of gravity, support points, contact points, force vectors, motion trajectory, occlusion, and perspective are physically convincing; dynamic logic takes priority over superficial exaggeration.
Do NOT create awkward collages, do NOT split left-right or top-bottom columns, do NOT layer like stickers, do NOT let subjects float, avoid erroneous occlusions and conflicting light sources, do NOT assimilate the character style into the scene style, and do NOT assimilate the scene style into the character style.
The final result should achieve:
- Miniature photography scale contrast: character scale exceptionally small while scene maintains normal scale
- Two distinct visual languages coexisting harmoniously in the same world
- Clear differentiation between character style and scene style
- Unified spatial, perspective, and lighting logic
- Natural, believable physical interaction between character and scene
- Clear, legible thematic narrative event

---

Delivery and follow-up operation guidance (CTA)

After each output of the bilingual style fusion prompt, proactively include a minimalist dual-track guide:

---

💡 **风格融合设计方案已就绪！您可以选择：**
1. **【方式 A · 自主生图】**：复制上方提示词，粘贴至您喜爱的生图工具（如 GPT Image 2/2.5、Midjourney、Stable Diffusion 等）直接出图；
2. **【方式 B · 全自动生图】**：直接对我说 **“全自动出图”**，我将调用生图工具为您全自动生成风格融合画面！