How to Use NanoBananaPro to Generate and Publish WeChat Spring Festival Red Envelope Covers

1. Prompt preparation

The prompt must satisfy WeChat Red Envelope cover constraints: canvas 957×1278 px, 3:4 ratio, RGB, 72 dpi, dominant "red envelope" background, top 25 % clean for white text, core visual elements in middle‑lower 50 %, bottom 15 % free of key elements. Example prompt (Chinese):

请生成一张微信红包封面图，画布尺寸 957×1278px，比例 3:4，RGB，72dpi。
主题/风格：关于2026年春节、有马元素、有春节孩子一起吃年夜饭的场景、有夜景放炮，整体人物的风格为海贼王风格，不是海贼王中的角色。
构图规则（强制）:
1) 画面上方25%区域避免复杂纹样与高频细节，保证白色文字清晰可读。
2) 核心视觉元素集中在中下部50%，形成明确焦点。
3) 底部15%避免放置关键元素。
设计要求（强制）：画面精致、有层次、细节清晰；不包含任何文字/标语/数字；不包含二维码；不包含水印、署名、logo；不出现边框式海报排版。
背景与装饰：在"红包红"主背景色上叠加[背景设计]（可为轻微纹理/渐变/金粉颗粒，但不能稀释主红色的占比），整体喜庆但不杂乱。
输出：高质量、干净画面。

2. Image generation via NanoBananaPro

The prompt is sent to the NanoBananaPro API through its web UI at https://imagen.apiyi.com/. The service returns a generated cover image that meets the visual constraints.

3. Custom Skill to automate generation

A Skill named nano-banana-pro-image-gen wraps the NanoBananaPro API call. Install it with:

npx skills add https://github.com/wuchubuzai2018/expert-skills-hub --skill nano-banana-pro-image-gen

After installation the Skill can be invoked via natural‑language commands to generate the cover image automatically.

4. Resizing the image with the image-resizer Skill

The generated image must be resized to the exact dimensions required by WeChat. Install the image-resizer Skill:

npx skills add https://github.com/wuchubuzai2018/expert-skills-hub --skill image-resizer

Core functions include pixel‑based resizing, percentage scaling, maximum‑size limiting, aspect‑ratio cropping, and compression to a target file size.

Example command‑line usages:

### Example 1: resize to specific dimensions
node resize_image.js input.png -w 800 -h 600 -o output.png

### Example 2: scale by percentage
node resize_image.js input.png -s 0.5 -o output.png

### Example 3: compress to target size (KB)
node resize_image.js input.png -S 500 -o output.jpg

### Example 4: crop to aspect ratio
node resize_image.js input.png -a 16:9 -o output.png

### Example 5: limit maximum dimensions while preserving ratio
node resize_image.js input.png --max-width 1920 --max-height 1080 -o output.png

Directory layout of the Skill:

image-resizer/
├── SKILL.md            # Skill main file
└── scripts/
    ├── resize_image.js # Image processing script
    └── package.json    # Dependency configuration

5. Uploading and publishing on the WeChat Red Envelope platform

After resizing, the image is uploaded through the WeChat Red Envelope Open Platform. The workflow is:

Log in to the platform and register an account.

Upload the resized cover image; the platform validates size and composition constraints.

After platform approval, the cover becomes available for purchase or exchange.

In the example, 2,000 cover units were exchanged.

6. Summary

The process demonstrates an end‑to‑end engineering workflow: craft a constraint‑aware prompt, generate a cover image with NanoBananaPro, automate generation via a custom Skill, resize the image to exact WeChat specifications using the image-resizer Skill, and finally upload and publish the cover on the WeChat Red Envelope platform.