Creating WeChat Stickers with NanoBananaPro: AI Image Generation Part 3
This tutorial walks through using Google’s Gemini‑based NanoBananaPro model to generate LINE‑style Q‑version stickers, split the images with an online tool, and submit them to the WeChat Sticker Open Platform, covering prompt design, copyright considerations, JSON specifications, and review feedback.
Character generation
Open the Gemini interface and generate a cartoon portrait for the base character of the sticker set. The example uses the well‑known character Han Li from “凡人修仙传”. The generated portrait is shown in the following images:
If a Gemini account is unavailable, the free service at https://www.miaoda.cn/?keyword=nano can be used to access the NanoBananaPro model.
Prompt creation
A detailed prompt requests a LINE‑style Q‑version sticker pack with a 4 × 6 layout, handwritten Simplified Chinese text, 2K resolution and a 16:9 aspect ratio. The prompt is:
紧接上图的头像形象,为我生成LINE风格的Q版表情包,注意衣服风格要正确,使用4*6布局,涵盖各种各样的聊天语句,适合在微信群聊中聊天使用。所有标准为手写简体中文,生成的图片需为2K分辨率,尺寸为16:9Submitting this prompt yields a full sticker sheet, which is displayed in the article.
Sticker sheet generation
The model returns a single image containing 24 stickers. Clear and specific prompts lead to higher‑quality results.
Image splitting
To obtain individual stickers, the sheet is uploaded to an online splitter built on Google tools (URL: https://gemini.google.com/share/91451e9895e9). A custom splitter with horizontal and vertical offset controls is also available at https://vibe.aijavapro.cn/#/meme-splitter.
After clicking the “Start Split” button, the tool downloads 24 separate PNG files, each representing a single sticker.
WeChat Sticker Open Platform submission
Register a personal account at https://sticker.weixin.qq.com, then upload the 24 individual images. Required metadata such as theme and style must be filled in.
First review failure
The initial submission is rejected. The rejection notice indicates that the stickers do not meet the platform’s technical requirements.
Structured JSON specification
To address the feedback, a text model is used to generate a structured JSON specification that defines the theme, style, technical requirements, and caption texts for each of the 24 stickers. The JSON is:
{
"theme": "马年Q版大头表情包 (微信审核优化版)",
"style": "LINE风格, 简约粗线条, 色彩鲜艳, 喜庆氛围",
"technical_requirements": {
"layout": "4列 * 6行 (共24个)",
"character_focus": "仅显示马的正面头部,严禁出现脖子以下身体、手臂或腿部",
"background": "浅米黄色或淡粉色(非纯白),营造活泼氛围",
"text_style": "手写体简体中文,位于头像下方或侧边"
},
"sticker_grid": [
["新年快乐 (灿烂笑脸)", "恭喜发财 (戴财神帽)", "万事如意 (wink眨眼)", "大吉大利 (头顶橘子)"],
["收到 (点头示意)", "没问题 (自信微笑)", "谢谢 (爱心眼)", "比心 (脸上红晕)"],
["抢红包啦 (双眼放光)", "谢谢老板 (崇拜眼神)", "错过一个亿 (流泪大哭)", "坐等红包 (乖巧期待)"],
["哈哈 (张嘴大笑)", "点赞 (开心眯眼)", "厉害了 (星星眼)", "太棒了 (撒花特效)"],
["在吗 (探头疑惑)", "问号 (歪头不解)", "吃惊 (圆睁双眼)", "害羞 (捂脸红晕)"],
["吃瓜 (呆萌注视)", "潜水 (冒泡泡)", "晚安 (闭眼睡颜)", "再见 (挥手感,仅露马蹄尖)" ]
]
}Regeneration and approval
Using the refined specification, the model generates a new sticker sheet with improved compliance. After splitting the sheet again, the 24 stickers are resubmitted.
The second review passes, and the approval notice is captured.
Final distribution
A QR code links to the published sticker pack, allowing users to download and use the stickers.
Signed-in readers can open the original source through BestHub's protected redirect.
This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactand we will review it promptly.
Ubiquitous Tech
A ubiquitous public account for pirate enthusiasts, regularly sharing curated experiences, tech learning, and growth insights. Currently publishing articles on AI RAG customer service, AI MCP technology, and open-source design. Personal free Knowledge Planet: Awakening New World Programmer.
How this landed with the community
Was this worth your time?
0 Comments
Thoughtful readers leave field notes, pushback, and hard-won operational detail here.
