图片转提示词生成器

上传参考图,生成可编辑、可复制、可复用的提示词

上传图片
上传图片,或从你的素材中选择
点击上传或选择图片
PNG / JPG / WEBP
上传或选择图片
输出格式
参考图片
生成提示词

A realistic vertical smartphone selfie of a young adult woman in her early 20s, shot indoors at night in a modern office or studio room. She is framed in a tight close-up from the upper chest upward, looking directly into the front camera with calm, slightly distant eyes and a soft closed-mouth expression. Her head is slightly tilted forward, and one hand is placed under her chin in a gentle “flower pose,” palm supporting the jawline. The other arm extends toward the camera, creating subtle selfie perspective distortion and a casual, intimate feeling. She has long, glossy black hair with a soft center part, smooth straight lengths falling over both shoulders, with a few natural flyaway strands catching the overhead light. Her makeup follows a refined contemporary East Asian / K-pop-inspired beauty style: clean fair neutral-cool skin, softly blurred base makeup, defined but delicate brows, long curled lashes, subtle winged eyeliner, faint lower-eye shimmer, grey-brown contact-lens-like eyes, soft pink blush, and glossy rose-pink lips. Preserve the original face feeling: oval face, large reflective eyes, slim nose bridge, small softly shaped lips, delicate chin. Visible skin includes the full face, neck area, both forearms, part of the upper arms due to the short sleeves, and the hand touching the chin. No waist, legs, hips, back, or chest exposure is visible. The skin tone is fair with a neutral-cool cast under cool indoor lighting. The skin should look smooth, hydrated, and softly satin to the touch, like lightly powdered but still supple skin, with gentle highlights on the nose bridge, cheekbones, forehead, and the back of the hand. She wears a fitted black short-sleeve T-shirt and silver accessories: thin hoop earrings, a chunky silver chain necklace, metallic rings, and a silver bracelet or cuff near the lower edge of the frame. The background is a real nighttime office interior: large dark windows with faint city lights outside, white ceiling panels, cool fluorescent or LED ceiling reflections, a desk with papers on the left, and a light wood cabinet or podium on the right. The lighting is cool and physical: overhead rectangular office lights create bright streaks on the top of her hair, while soft front-facing smartphone/room light fills her face evenly. The background remains visible but slightly less emphasized. The image should feel like it was taken late at night after work, quiet and slightly tired but composed, as if she had paused for one polished selfie before leaving. Keep the framing slightly imperfect and human: mild high-angle selfie perspective, close lens distance, natural facial asymmetry, tiny hair flyaways, real office details, slight smartphone smoothing, and subtle digital noise. Avoid making it look like a studio fashion shoot, an AI-perfect portrait, or an overly airbrushed commercial beauty image.

它会从图片里读出什么?

它不只看主体,也会提取风格、光线、构图、镜头感和影响画面质感的小细节,方便你更接近原图效果。

主体和场景

识别主要人物、物体、动作、环境,以及它们之间的关系。

风格和媒介

判断画面更接近摄影、电影感、插画、3D、动漫、商品图还是海报。

光线和色彩

提取光源方向、明暗对比、色彩组合、阴影和整体情绪。

构图和镜头

把取景、视角、镜头感、景深、透视和裁切方式写进提示词。

细节和氛围

补充材质、服装、纹理、环境、情绪和画面完成度。

可编辑提示词

输出可继续修改的提示词,可用于 Midjourney、GPT image、Nano Banana 等模型。

Promptsref 优势不只是生成一句图片描述

我们会把图片当成一份提示词说明来解析,同时给出详细的 JSON 提示词和更短的自然语言提示词。

01

结构化 JSON 提示词

JSON 版本会拆出主体、姿势、服装、光线、镜头、背景、氛围、限制条件和反向提示词。信息越清楚,模型越容易复刻原图的视觉效果。

02

同时给出短提示词

除了详细 JSON,我们也会给一版自然语言提示词。它更短,更适合快速复制、轻量修改和直接生成。

03

支持多语言输出

解析前可以选择输出语言。图片细节会保留下来,但提示词会用你习惯的语言表达。

Reference portrait paired with a shortened structured JSON prompt

怎么使用图片转提示词

01 · Reference image

Archival-style reference photograph of an elderly shopkeeper with his dog

02 · Turn image to prompt

A vertical, documentary-style street photograph with a strong early-20th-century atmosphere: an elderly white-bearded man stands in the narrow doorway of an old dark shop or pub, accompanied by a small scruffy cream-colored terrier-like dog. The man is centered but slightly off-axis, leaning casually against the doorframe with one leg crossed over the other and his weight settled into the threshold. He wears a weathered brown tweed flat cap, a dark charcoal-gray overcoat hanging open, a textured gray herringbone waistcoat, a muted rust-brown tie, loose high-waisted gray trousers, and worn dark leather shoes. His left hand rests inside a trouser pocket while his right hand raises a short smoking pipe to his mouth. His expression is quiet, stern, and contemplative, with narrowed eyes directed toward the camera and a face marked by deep age lines, a full white mustache, and a neatly heavy white beard. Smoke curls softly from the pipe toward the shadowed doorway. The dog sits in the lower-left foreground, looking upward toward the man with an alert, affectionate expression. Its fur is shaggy, wiry, and uneven, pale ivory with slightly darker beige patches around the ears and muzzle; the fur should appear tactile, dusty, and naturally tousled rather than groomed. The man’s face and hands show dry, weathered skin with visible wrinkles, creases, and realistic texture, softly touched by warm reflected light. No other body skin is visible. Frame the scene as a close full-body portrait from a slightly low, eye-level-to-chest-height viewpoint, as if taken by a photographer standing just outside the doorway with a vintage 35mm camera and a normal-to-slightly-wide lens around 40mm. Use a vertical 4:5 composition. Keep the man dominant in the center, the dog anchoring the lower-left corner, and the doorway creating strong dark vertical framing on both sides. Allow slight asymmetry, imperfect centering, subtle lens distortion, a little foreground obstruction, and the feeling of a found archival photograph rather than a polished studio portrait. The old shopfront is cramped and heavily worn: dark brown and black wooden doorframes, chipped paint, stained surfaces, faded hand-painted signs, shelves of bottles, stacked papers, tins, and small miscellaneous goods visible along the edges. The interior behind the man is almost black, creating a deep recess that makes his face, clothing, pipe smoke, and the dog stand out. Use warm, low-contrast daylight mixed with amber reflections from aged wood and nearby shop surfaces. A soft directional light enters from the front-left, grazing the man’s beard, cap, coat shoulders, waistcoat texture, and the dog’s fur while leaving the doorway behind him in dense shadow. Add faint cool ambient fill from the street, subtle rim light on the coat edge, and naturally uneven exposure across the frame. The light should behave physically: highlights catch on worn fabric fibers, the pipe, dusty bottles, and rough wood; shadows remain brown-black and slightly lifted rather than pure digital black. Use shallow-to-moderate depth of field, keeping the man’s face, pipe, waistcoat, and dog recognizable while allowing the shop edges and background objects to soften gradually. Create a tactile analog image with muted sepia, tobacco brown, olive gray, charcoal, faded cream, and subdued rust tones. Include fine film grain, slight halation around bright edges, soft focus falloff, minor dust specks, faint scratches, gentle vignette, restrained color fading, and the imperfect tonal softness of an aged color photograph or hand-tinted archival print. Avoid a modern digital-clean appearance, fashion-editorial polish, exaggerated sharpness, glamour retouching, plastic skin, contemporary clothing, excessive smoke, dramatic cinematic lighting, or a pristine background. The image should feel like a fleeting moment that happened decades ago and was discovered later: a tired shopkeeper pausing in the doorway, his loyal dog waiting beside him, while the street continues just outside the frame. Caption energy: “He had been standing there all afternoon, pipe glowing quietly, the dog watching the world go by.”

03 · Generated image

Generated archival-style portrait of an elderly shopkeeper with a terrier
📱

上传或选择要分析的图片

等待几秒钟生成提示词

😊

复制提示词,或先修改再使用

什么时候适合用图片转提示词

01

拆解一张参考图的风格和画面结构。

02

把人像、商品图、海报或插画整理成可复用提示词。

03

把视觉灵感转换成 GPT image、Nano Banana、Midjourney 等模型可用的 prompt。

04

学习高质量图片提示词如何描述主体、风格、光线和构图。

05

在保持同一视觉方向的前提下,扩展更多 prompt 变体。

Reference street photo paired with a short natural-language prompt

生成后可用于不同模型

保留视觉描述,再按你使用的模型补充比例、参数或风格语法。

GPT image 2.0

适合用自然语言描述主体、姿势、镜头、光线和画面质感。

Nano banana

适合参考图复刻、视觉改写,以及需要理解图片细节的场景。

Midjourney

适合风格复刻、电影感参考、SREF 探索和概念图生成。

其他 AI 生成模型

把提取出的提示词作为基础,再补充对应模型的参数或语法。

Reference anime artwork paired with a Chinese-language prompt

图像转提示词工具可以将图像转换为描述性提示词。它使用人工智能分析图像内容并生成相应的提示词,这些提示词可用于创建类似图像或其他文本到图像的生成任务。该工具帮助用户更好地理解和描述图像内容,促进AI图像生成和创意工作。

你上传的图片会被安全保存,用于生成提示词,并方便你之后查看结果和历史记录。图片默认仅对你的账号可见,其他用户无法查看,也不会出现在公开画廊中。

图像转提示词在AI图像生成中扮演着至关重要的角色,它帮助用户从现有图像中提取关键信息并生成准确的提示词。这些提示词可以作为创建类似或改进图像的新描述,帮助用户更好地理解AI如何"看待"图像,并提高他们的提示词写作技巧。

图像转提示词输出通常包括对主体、场景、艺术风格、色彩方案、光照条件和整体情绪或氛围的描述。它还可能描述图像中存在的特定主题、动作、姿势、构图、视角、着色、风格或详细特征,旨在提供全面的文本表示。

使用图像转提示词很简单:只需将图像上传到我们的工具。我们的AI将分析图像并生成详细的提示词。然后,您可以使用这个提示词作为使用AI图像生成器创建新图像的起点,或根据您的特定需求进一步完善它。

我们的图像转提示词生成器使用先进的AI模型提供高度准确的图像描述。然而,像所有AI技术一样,它有时可能会误解某些元素。我们不断努力提高其准确性,并鼓励用户根据其特定目的审查和调整生成的提示词。

我们的图像转提示词生成器设计用于处理各种类型的图像,包括照片、数字艺术和插图。然而,它在清晰、高分辨率图像上表现最佳。一些高度抽象或复杂的图像可能会导致更一般化的描述。

使用图像转提示词可以帮助您了解AI如何解释视觉信息。这有助于您理解在生成图像时在自己的提示词中包含哪些元素,并作为学习有效图像提示词的词汇和结构的绝佳工具。

不,图像转提示词生成器不允许任何NSFW内容。上传NSFW内容将被我们的系统阻止。