OpenZhenZhen AI 提示词画廊 · gpt-image-2.5 · image

打字机上的诗歌地形:文字构建的微缩世界

Poetic Terrain on a Typewriter: A Miniature World Built of Words

2×2微距编辑网格:四台复古打字机将用户提供或授权的短诗摘录转化为从纸面升起的3D地形,配以诗意物件、微型雕像、情绪化灯光和黄铜元数据键。

打字机上的诗歌地形:文字构建的微缩世界 — AI image creation
打字机上的诗歌地形:文字构建的微缩世界

提示词 / Prompt

中文提示词

2x2 微距摄影编辑网格,主题为“打字机输出的 3D 地形”。使用一段由用户提供、表达思念的短诗作为创作输入;只使用用户提供或已获授权的短诗摘录,不生成或复现未提供的长篇受版权保护文本。展示四台分别对应不同诗歌情绪的复古打字机,每台滚筒中装有一张纸;打出的短摘录从纸面物理升起,形成 3D 地形,关键短语如石刻般嵌入地表。按照微型玫瑰、时钟、门和纸船等元素,在地形中摆放微型物件。加入一尊无法辨认身份的微型文学雕像,让它漫步于文字景观中,不呈现任何可识别的真实诗人肖像。灯光随情绪变化:忧郁时用薄雾,愤怒时用较硬的光,思念时用柔和金色;具体表现为忧郁配薄雾、愤怒配更硬朗的光线、思念配柔和金光。每个面板都放置一枚黄铜打字机键,显示清晰易读的简洁元数据“Poet | Collection | Year”,填入 Anonymous Poet | Small Collection | 2026。背景加入符合时代氛围但不带品牌的物件。保持四格对齐,使用浅景深、精致的纸张与金属质感,并让文字呈现从纸面升起的空间效果。排除标志、水印、平台名称及未经授权的长篇诗句。

English prompt

2x2 macro-photography editorial grid themed "Typewriter Output as 3D Landscapes". Use a user-provided short poem about longing as the creative input and only use short user-provided or licensed poem excerpts; do not generate or reproduce long unprovided copyrighted text. Show four vintage typewriters associated with different poetic moods, each with one sheet loaded in the roller. The typed short excerpt physically rises from the page into a 3D topographical landscape, with key short phrases engraved into the terrain like carved stone. Place miniature objects such as a boat, rose, clock or door according to a tiny rose, clock, door and paper boat. Include an unidentifiable miniature literary figurine wandering through the word landscape rather than a recognizable real poet likeness. Shift lighting with fog for melancholy, hard light for anger, soft gold for longing: fog for melancholy, harder light for anger and soft gold for longing. Each panel has a brass typewriter key with clean legible metadata reading "Poet | Collection | Year", populated from Anonymous Poet | Small Collection | 2026. Add era-relevant but unbranded background objects. Keep the four panels aligned, use shallow depth of field, refined paper and metal textures, and make the words visibly rise from the page. Exclude logos, watermarks, platform names and unauthorized long poem text.

查看 API 模型

探索更多图片和视频提示词 · 生成图片 · 生成视频 · 在 AI 画布中创作

  • 黄昏

    A quiet unposed fine art photograph of a young East Asian woman curled sideways in a wicker chair by the window at dusk, caught in a private unguarded moment, not posing. Long wavy dark brown hair falling loose over one shoulder and across her face, slightly tousled. Her gaze is lowered and averted

  • 热血少年宝玉角色设定图

    以当前固定参考肖像作为预览人物槽,为角色名“宝玉”制作少年漫风格的虚构示意设定图。面部与发型依据参考中可见特征;服装、姿势、配色和类型化光效均为可编辑的创作默认值,不代表对用户个人资料或性格的了解。 Create a fictional illustrative shonen-style character sheet named “宝玉,” using the current fixed portrait as the preview-subject slot. Base face and hair only on visible reference features; costume, p

  • 黄昏藤椅

    A quiet unposed fine art photograph of a young East Asian woman curled sideways in a wicker chair by the window at dusk, caught in a private unguarded moment, not posing. Long wavy dark brown hair falling loose over one shoulder and across her face, slightly tousled. Her gaze is lowered and averted

  • 薰衣草墙花卉人像

    无需外部参考,生成成年女性越过薰衣草色墙垂双臂持大花束的3:4写实时尚画面。

  • 极简文字的个人色彩诊断图

    以固定参考肖像为对象,通过服装色彩样片进行图表优先的视觉比较,文字简短,不写段落。结果是基于可见肖像的造型建议,不宣称经过校准的科学诊断。 Use the configured portrait as the subject and compare clothing-color swatches in a diagram-first layout. Keep labels brief and omit paragraphs. Present the result as a visual styling suggestion based on the visible portrait, not

  • 古筝弹唱古典美人

    一位成年表演者在明确标注为虚构演绎的直播间中扮演历史人物陈圆圆,弹奏古筝并演唱古风歌曲;固定参考仅用于表演者的面部身份,不代表陈圆圆的真实长相。弹幕保留原文“圆圆唱一曲”。 An adult performer portrays the historical Chen Yuanyuan in a clearly fictionalized livestream concept, playing the guzheng and singing a classical-style song. The fixed reference anchors only the performer’s face