选一个场景 → 描述这对情侣 → 选风格 → 调美学参数 → 选尺寸 → 生成高级写真指令。
复制后粘贴到 即梦 / 通义万相 / ChatGPT / Gemini；配两人正脸照作参考更还原。
情侣写真是一个免费的 AI 情侣/婚纱写真提示词生成器，帮情侣、婚礼策划和写真工作室快速产出照片级的双人大片指令。内置 59 个场景，分 5 大类（韩系棚拍、高级编辑、都市街拍、浪漫旅拍、环球史诗婚纱），可调颗粒、色温、景深与面部相似度。用法：选一个场景 → 调滑杆并写清服装/氛围 → 复制生成的提示词，粘贴进即梦、Nano Banana 或 Midjourney 即可出图。
Couple Studio is a free AI couple- and wedding-portrait prompt generator for couples, wedding planners and photo studios who want photoreal two-person shots fast. It ships 59 scenes across 5 groups (K-studio, high editorial, streetwear, city travel, epic destination wedding) with sliders for grain, color temperature, depth of field and face likeness. Pick a scene, tune the sliders and describe the outfit and mood, then copy the prompt into Jimeng, Nano Banana or Midjourney to render it.
完全免费，无需注册或登录。打开页面选场景、调滑杆就能生成提示词，没有次数限制。
It is completely free with no signup or login required. Open the page, pick a scene, tune the sliders and generate — there is no usage limit.
提示词是通用的文生图 prompt，可以粘贴进 ChatGPT / DALL·E、Midjourney、即梦、Nano Banana、Gemini 等主流生图工具使用，中英双语都能识别。
It is a general-purpose text-to-image prompt you can paste into ChatGPT / DALL·E, Midjourney, Jimeng, Nano Banana, Gemini and other mainstream image generators. Both the Chinese and English text are understood.
纯文字提示词本身无法复刻真实人脸，只会画出「相似的人」。要还原你们的五官，必须把清晰正脸照上传到支持人物参考的工具或模式里，再粘贴本站的提示词——例如即梦的「人物参考/上传形象」、Nano Banana 的上传参考图、Midjourney 的 --cref。光线均匀、单人、正脸的参考照效果最好。
A text-only prompt cannot replicate a real face — it only draws a similar-looking person. To capture your actual features you must upload clear front-facing photos into a tool or mode that supports character reference, then paste this prompt — for example Jimeng's character-reference / upload-portrait mode, Nano Banana's reference-image upload, or Midjourney's --cref. Evenly lit, single-person, front-facing photos work best.
输出的是提示词（prompt），本站不直接生成图片。你把提示词复制到即梦、Nano Banana 或 Midjourney 等生图工具里，由它们出图。
You get a prompt — this site does not generate images directly. Copy the prompt into an image generator such as Jimeng, Nano Banana or Midjourney and it produces the image.