生成描述之后有一个 复制,有一个下载 JSON 。
你面向的用户是谁啊?普通用户的话,你让普通用户下载 JSON 干啥?码农的话,我用你这玩意儿干啥?
你给我生成出来的是啥文本格式啊?
```
Here are a few options for a catchy caption, choose the one that best fits your desired tone!
Option 1 (Sweet & Simple): Sweet and serene, radiating youthful charm in her classic sailor uniform. Pure school day vibes!
Option 2 (Reflective): A gentle smile and a moment of calm reflection. There's a timeless beauty in youthful innocence and a crisp school uniform.
Option 3 (Engaging): Fresh faces and classic uniforms never go out of style! Loving this sweet schoolgirl aesthetic.
Appropriate Hashtags: #SchoolGirl #SailorUniform #Seifuku #YouthfulVibes #SweetSmile #StudentLife #AsianBeauty #Portrait #Innocence #FreshFace #UniformStyle #SchoolDays #CuteGirl #GentleLook
```
能不能搞搞 prompt engineering ,让 LLM 直接生成结构化的数据,你自己先 parse 好,最后把干干净净的字给用户。最后给用户一句话、几个 hashtag ,用户点一下「复制」就能利利索索的复制走,直接贴到 Instagram 里?
像这样:
