SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillshuashu-wechat-image

huashu-wechat-image

Design & creation

View repository

Updated 10-15 16:37

About this skill

WeChat Official Account Image Workflow

⚠️ Core Principle: Propose First, Generate Later

Absolutely never skip the design proposal and generate images directly. Correct process:

理解内容 → 设计提案（2-3个方向）→ 用户选择 → 生成 → 预览确认 → 上传

For design aesthetic profiles and proposal formats, see the sister skill xhs-image/SKILL.md, which also applies to WeChat official account images.

Core Parameters

Image Type Size Ratio Playwright viewport AI Prompt Aspect Ratio Declaration
Headline Cover 1800 x 766 px 2.35:1 --viewport-size=1800,766 2.35:1 ultra-wide landscape, 1800x766 pixels
Body Wide Image 1920 x 1080 px 16:9 --viewport-size=1920,1080 16:9 landscape, 1920x1080 pixels
Body Square Image 1440 x 1080 px 4:3 --viewport-size=1440,1080 4:3 landscape, 1440x1080 pixels
Infographic 1080 x Free Free --viewport-size=1080,N Determined by content

Step 0: Determine Image Requirements ✅ User Confirmation

Show options to the user and wait for confirmation:

Type Description Number of Images Recommended Path
A. Cover Only Headline cover image 1 image AI generation
B. Body Images Section illustrations to aid reading comprehension 3-8 images Mainly AI generation
C. Full Set of Images Cover + all body images 4-10 images Mainly AI generation
D. Pure Infographics Data comparisons, flowcharts, checklists 1-5 images AI generation (use HTML as fallback for precise data tables)

Ask the user：

  1. Do you want a cover, body illustrations, or the full set?
  2. What is the article path/content? (Analyze the content to determine image placement)
  3. How many images are needed in total?

Image Quantity Recommendations

Article Length Recommended Number of Images Includes
< 1500 words 2-3 images Cover + 1-2 body images
1500-3000 words 4-6 images Cover + 1 image per core section
3000-5000 words 6-8 images Cover + section images + infographics
> 5000 words 8-10 images No more than 10 images, to avoid excessive interruption

Step 1: Style Selection ✅ User Confirmation

Based on the user's article type, recommend 3 styles for selection.

Automatically recommended by article type

Article Type First Recommendation Second Recommendation Third Recommendation
AI Tool Review Minimalist Professional Editorial Magazine Data Infographic
Technical Tutorial Minimalist Professional Hand-drawn Whiteboard Editorial Magazine
Product Launch Editorial Magazine Big-Character Poster Minimalist Professional
In-Depth Analysis Editorial Magazine Data Infographic Minimalist Professional
Personal Story Snoopy Warm Comic Warm Narrative Editorial Magazine
Industry Observation Big-Character Poster Editorial Magazine Data Infographic

Show the user 3 recommended styles, each including:

  • Style name + one-sentence description
  • Use cases
  • Color tendency

Ask the user: Which style do you want? Or do you have your own reference in mind?

Complete style library: references/style-gallery.md

Step 2: Choose generation path ✅ User confirms

Recommend a path based on content characteristics, and show the comparison to the user:

Path A: HTML → Screenshot Path B: AI Generation
Text accuracy 100% (code-controlled) Chinese may have errors (needs verification)
Layout control Pixel-level precision AI free-form
Visual creativity Medium (depends on design ability) High (AI is creative)
API cost Zero (purely local) Each image consumes Gemini API
Speed Fast (a few seconds) Slow (10-30 seconds/image)
Suitable for Lots of text, lots of data, lists, infographics Covers, atmosphere images, creative illustrations
Editability Just modify the HTML and take a new screenshot Needs to be regenerated
Dark mode CSS directly controls the color scheme Must be declared in the prompt

⚠️ Huashu preference: AI generation first

HTML screenshots look too flat, like PowerPoint templates, lacking texture and design sense. AI-generated images have rich visual details, and the quality far exceeds HTML. By default, use the AI generation path. Only use HTML as a fallback for "complex data tables that must be verbatim accurate."

Path recommendation rules

Image type Recommended path Reason
Headline cover AI Strong visual impact, good texture
Body atmosphere/scene image AI Highly creative
Body infographic AI AI can create infographics with a sense of design
Flowchart/step diagram AI AI can add visual hierarchy
Comparison chart AI Higher visual quality
Precise data table (10+ cells) HTML The only scenario that needs HTML: large amounts of precise numbers

By default, recommend the AI path; no longer ask about path selection each time.

Step 3: Generate image

Path A: HTML → Playwright Screenshot

3-A-1. Create HTML

Based on the selected style and the content provided by the user, generate an HTML file.

HTML template requirements:

  • Canvas size: choose according to image type (cover 1800x766, body wide image 1920x1080, body square image 1440x1080)
  • Font:font-family: "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei", sans-serif
  • Dark mode friendly: use #F5F5F5 instead of pure white, use #1A1A2E instead of pure black
  • Cover safe zone: core information concentrated in the central square area (766x766)

Cover big-text poster template example (2.35:1):

<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<style>
  * { margin: 0; padding: 0; box-sizing: border-box; }
  body {
    width: 1800px; height: 766px;
    background: #1A1A2E;
    display: flex; flex-direction: column;
    justify-content: center; align-items: center;
    padding: 60px 400px; /* 左右大padding确保内容在安全区 */
    font-family: "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei", sans-serif;
  }
  .title {
    font-size: 96px; font-weight: 900;
    color: #FFFFFF; text-align: center;
    line-height: 1.3; letter-spacing: 4px;
  }
  .subtitle {
    font-size: 36px; font-weight: 400;
    color: #E2B714; text-align: center;
    margin-top: 30px;
  }
  .accent-line {
    width: 200px; height: 4px;
    background: #E2B714;
    margin: 30px auto;
  }
</style>
</head>
<body>
  <div class="title">Claude vs Codex</div>
  <div class="accent-line"></div>
  <div class="subtitle">重度用户的20分钟实战对比</div>
</body>
</html>

Body infographic template example (4:3):

<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<style>
  * { margin: 0; padding: 0; box-sizing: border-box; }
  body {
    width: 1440px; height: 1080px;
    background: #F5F5F5;
    padding: 60px 80px;
    font-family: "PingFang SC", "Hiragino Sans GB", "Microsoft YaHei", sans-serif;
  }
  .heading {
    font-size: 56px; font-weight: 800;
    color: #1A1A1A; margin-bottom: 40px;
    line-height: 1.3;
  }
  .item {
    display: flex; align-items: flex-start;
    margin-bottom: 28px; gap: 20px;
  }
  .num {
    font-size: 40px; font-weight: 900;
    color: #4A90D9; min-width: 50px;
  }
  .text {
    font-size: 32px; color: #333;
    line-height: 1.5;
  }
  .highlight { color: #FF6B35; font-weight: 700; }
</style>
</head>
<body>
  <div class="heading">Claude Code 五大优势</div>
  <div class="item"><span class="num">01</span><span class="text"><span class="highlight">Agent模式</span>自主规划执行复杂任务</span></div>
  <div class="item"><span class="num">02</span><span class="text">原生<span class="highlight">终端集成</span>直接操作文件系统</span></div>
  <div class="item"><span class="num">03</span><span class="text"><span class="highlight">多文件编辑</span>一次修改整个项目</span></div>
</body>
</html>

3-A-2. Screenshot

npx playwright screenshot "file:///path/to/card.html" output.png \
  --viewport-size=[宽],[高] --wait-for-timeout=1000
  • Cover:--viewport-size=1800,766
  • Body wide image:--viewport-size=1920,1080
  • Body square image:--viewport-size=1440,1080

3-A-3. ✅ Preview confirmation

Use the Read tool to show the screenshot result to the user.

Ask the user：

  • Is the text content correct?
  • Are you satisfied with the layout and color scheme?
  • Is the cover's core information within the safe zone? (Cover image only)
  • What needs to be adjusted?

If adjustment is needed: modify HTML → take screenshot again → confirm again. (The iteration cost of the HTML path is extremely low)

Path B: AI generation (Gemini 3 Pro Image)

3-B-1. Build Prompt

Cover Base Style Prompt:

[Base Style]:
VISUAL REFERENCE: [一句话描述具体风格美学]
CANVAS: 2.35:1 ultra-wide landscape, 1800x766 pixels, high quality rendering.
SAFE ZONE: Center square (766x766) contains all key text and visuals.
COLOR SYSTEM: [描述色彩情绪，不指定比例]
TEXT RENDERING: Chinese text must be large, clear, readable. Title in center safe zone.

Body Base Style Prompt:

[Base Style]:
VISUAL REFERENCE: [一句话描述具体风格美学]
CANVAS: [16:9 landscape, 1920x1080 pixels / 4:3 landscape, 1440x1080 pixels], high quality rendering.
COLOR SYSTEM: [描述色彩情绪]
DARK MODE: Use medium-tone backgrounds, avoid pure white (#FFFFFF) and pure black (#000000).

Per-Image Prompt：

Create a [style] [cover/illustration] for a WeChat article about [topic].

[Base Style]

DESIGN INTENT: [用户看到后应该产生什么情绪/行为]

TEXT TO RENDER:
- Title: "[标题，封面≤10字，正文可少或无]"
- [其他文字]

[1-2句画面情绪描述，让 AI 自由发挥构图]

3-B-2. ✅ Prompt confirmation

Show the prompt to be used to the user, and wait for confirmation or modification. Especially confirm:

  • Whether the text content to be rendered is correct
  • Whether the design intent is accurate
  • Whether there are any other requirements

3-B-3. Generate

export $(grep GEMINI_API_KEY ~/.claude/.env) && \
uv run /Users/alchain/Documents/写作/.claude/skills/wechat-image/scripts/generate_image.py \
  --prompt "[完整prompt]" \
  --filename "[timestamp]-wechat-[类型]-[描述].png" \
  --aspect [cover|wide|standard|square]
  • --aspect cover → 2.35:1 headline cover
  • --aspect wide → 16:9 body wide image (default)
  • --aspect standard → 4:3 body square image
  • --aspect square → 1:1 square image

Multiple images can be generated in parallel (run_in_background=true）。

3-B-4. ✅ Preview confirmation

Use the Read tool to show the generation result to the user.

Checklist:

  1. Is the text accurate? (No garbled characters or typos in Chinese)
  2. Is the ratio correct? (Cover 2.35:1, body 16:9 or 4:3)
  3. Cover safe zone: is the core information within the central square? (Cover only)
  4. Dark mode: has pure white background and pure black background been avoided?
  5. Are you satisfied with the style?

Ask the user: satisfied / regenerate / adjust prompt?

Path C: Hybrid path (recommended for full sets of illustrations)

Suitable for full sets of illustrations:Use AI generation for the cover (eye-catching), use HTML rendering for body infographics (precise information), use AI generation for body atmosphere images。

Execution order:

  1. First generate the cover according to Path B → user confirmation
  2. Extract a consistent style from the cover's color scheme/atmosphere
  3. Use Path B for body atmosphere images → user confirmation
  4. Use Path A for body infographics/data charts → user confirmation

Step 4: Upload to image hosting

python3 /Users/alchain/Documents/写作/tools/upload_image.py "[图片路径]"

Return the ImgBB permanent link.WeChat Official Account articles must use network links, local paths become invalid after publishing.

Quick reference

Cover safe area

┌─────────────────────────────────────┐
│         │ 766x766  │               │
│  装饰区  │ 安全区域  │  装饰区       │ 1800 x 766
│         │ （核心）  │               │
└─────────────────────────────────────┘
           ↑ 朋友圈裁切区域 ↑

Key points for dark mode adaptation

  • Use #F5F5F5 (light gray) instead of pure white for the background color
  • Use #1A1A2E (dark purple-gray) instead of pure black for the background color
  • Use #595959 or #3F3F3F for body text color, not pure black #000000
  • Avoid shadow effects and white borders
  • Use low-saturation color schemes

Chinese text rendering (unique to AI path)

  • Cover title ≤ 10 characters
  • Text on body images ≤ 20 characters/line
  • Rendering infographic text in HTML is more reliable
  • Must verify each image one by one

Font recommendations (unique to HTML path)

  • Title:"PingFang SC" Bold / "Microsoft YaHei" Bold
  • Body:"PingFang SC" Regular
  • Commercially safe alternatives: Alibaba PuHuiTi, Source Han Sans

Huashu Tech account color scheme

Scheme name Background Primary color Accent color Use case Dark mode
Warm Gray Professional #F5F0EB #D97706 #4A90D9 AI tools, sharing Good
Minimal Professional #F5F5F5 #4A90D9 #FF6B35 Tutorials, comparisons Medium
Midnight Gold #1A1A2E #E2B714 #FFFFFF Product launches Good
Terminal Green #1A1A1A #00FF41 #888888 Programming-related Good

Golden Rules

  • Place the cover's core information in the central square safe area
  • Use mid-tone backgrounds for body images (compatible with dark mode)
  • For infographics, prioritize HTML → screenshot (accurate text)
  • For atmospheric images, prioritize AI generation (highly creative)
  • Upload images to an image hosting service to get permanent links
  • Keep illustration style consistent within the same article
  • Include one image every 800-1200 characters
  • At least one image per H2 section

Image placement strategy

Position Necessity Type
Below the title (cover) Required Cover image / atmospheric image
After each H2 heading Recommended Section illustration
At data/comparison points Recommended Infographic
At product/tool introductions Optional Screenshot / AI concept image
Before the concluding summary Optional Closing illustration

Decision flowchart

用户需求 → Step 0 确定需求
                ↓
         Step 1 选风格（展示3个选项）
                ↓
         Step 2 默认AI生成（仅精确数据表格走HTML）
                ↓
         Step 3 生成 → 预览 → 用户确认
           ├→ 满意 → Step 4 上传
           ├→ 文字渲染有误 → 该张改用HTML兜底
           └→ 不满意 → 调整prompt重新生成

Related Skills

Skill Function
xhs-image Xiaohongshu illustrations (sister skill)
image-to-slides PPT illustrations (source of the style library)

Reference files

  • references/style-gallery.md — Complete style library and prompt templates
  • references/design-guidelines.md — WeChat Official Account platform design guidelines

Produced by Huashu | AI Native Coder · Indie Developer WeChat Official Account "Huashu" | 300k+ followers | AI tools and productivity Representative works: Xiaomao Fill Light (Top 1 on AppStore paid chart) · "One Book to Master DeepSeek"