Visual AI Prompt Editor Eliminates ‘Spell’ Anxiety, Tweaks Like Ordering Food

Magic demonstration

Case 1 – converting a long English prompt describing a “California girl” portrait into a concise Chinese module and generating an Asian‑styled image while preserving the original lighting, composition and editorial vibe.

# 原提示语
A high-resolution, close-up editorial-style portrait of a young woman standing against a textured gray concrete wall, captured with direct on-camera flash that creates a raw, moody, high-contrast aesthetic. ...

Original image:

Edited prompt (Chinese module):

# 修改的提示语
高分辨率 ，特写 编辑风格的肖像 ，描绘一位 年轻女性 站在 砖墙 前，以 直接机顶闪光灯 拍摄，营造出 原始、忧郁、高对比度的美学 … --ar 9:16

Resulting image shows the “California girl” style transferred to an Asian face while keeping the street‑editorial atmosphere.

Case 2 – modifying a JSON‑style prompt to replace the protagonist with a virtual assistant “Sienna” and change the setting to a ruin‑filled rooftop

# 原始提示语
{ "prompt": "Cinematic film still, **extreme top-down MEDIUM shot**. The woman is positioned at the center of the frame, **looking directly up into the camera lens with intense eye contact**. ...", "negative_prompt": "looking away, eyes closed, ...", "style": "high-fashion editorial, cinematic film still, analog film aesthetic", "camera": "Top-down medium shot, looking directly up at camera", "aspect_ratio": "3:4", "lighting": "natural daylight, high contrast, catchlights in eyes" }

Initial generation (skyscraper scene):

Edited prompt after visual configuration:

电影胶片风格 ，极致俯视  中景镜头 。这位 年轻女性，100%还原上传人物五官特征 位于画面中央，直视上方镜头 ，眼神 充满张力 … 废墟 … 8K写实主义 --ar 9:16

Resulting image portrays Sienna as a post‑apocalyptic heroine, retaining cinematic lighting and film‑grain texture.

Usage workflow

Behind the scenes the system performs multiple LLM calls and keyword matching to decompose the prompt.

Step 1 – Paste the original prompt

The editor accepts any format: JSON, Markdown, plain text, English, Chinese, or mixed.

Step 2 – Launch visual decomposition

The prompt is automatically split into structured fields displayed in an editable panel.

Each field provides bilingual suggestions (e.g., camera angle offers “top‑down, low‑angle, eye‑level”).

Selections update the prompt in real time.

电影胶片风格 ，极致俯视  中景镜头 。这位 年轻女性，100%还原上传人物五官特征 位于画面中央，直视上方镜头 ，眼神 充满张力 …

Step 3 – Export

The final prompt can be exported as structured JSON for integration into other pipelines.

# 提示语json格式
{
  "id": "prompt_1768224338957",
  "name": "Cinematic Urban Vertigo Portrait",
  "template": "{{art_style}} , {{lens_angle}}  {{lens_type}} . The {{character}}  is positioned at the center of the frame, {{pose}}  with {{expression}} . ...",
  "variables": { "art_style": { "value": "Cinematic film still", "options": ["Photographic portrait"] } }
}