wechat-article-to-markdown

WeChat Article Pipeline

A self-contained workflow skill: input a WeChat article link, output an organized Markdown article directory.

Use Cases

• The user provided an mp.weixin.qq.com article link
• Need to directly produce Markdown without manually chaining multiple skills
• Need to copy the skill to another project or share it with others

Usage

python scripts/wechat_article_pipeline.py <wechat_article_url> [--output-dir <目录>] [--save-html]

Example

python scripts/wechat_article_pipeline.py "https://mp.weixin.qq.com/s/xxxxxx" --output-dir "./articles"

If you also need to keep the cleaned HTML after scraping:

python scripts/wechat_article_pipeline.py "https://mp.weixin.qq.com/s/xxxxxx" --output-dir "./articles" --save-html

Workflow

Execute by completing the following steps:

1. Request the WeChat article page
2. Extract the title, author, official account name, and body area
3. Fix WeChat article lazy-loaded image URLs
4. Download body images to the local output directory
5. Convert the body to Markdown
6. First run the script's built-in Markdown formatter to clean up headings, blank lines, invalid links, missing images, and official account noise content
7. If the markdown-formatter skill exists in the current environment, automatically run markdown-formatter rule cleanup once more on the generated Markdown
8. Output the article directory, Markdown path, and processing summary

Output Structure

articles/
├── 01_文章标题/
│   ├── 文章标题.md
│   ├── source.html
│   ├── image_01.jpg
│   ├── image_02.png
│   └── ...

Notes:

• Directories automatically increment by number, for example 01_, 02_, 03_
• The Markdown file name stays consistent with the directory title
• source.html is only output additionally when using --save-html

Parameters

• <wechat_article_url>: WeChat article link
• --output-dir: Output directory, default: {工作区}/articles
• --workspace-dir: Workspace directory, automatically detected by default
• --save-html: Whether to save the cleaned HTML file
• --timeout: Request timeout in seconds, default 30

Workspace Detection

The script detects the workspace directory in the following priority order:

1. --workspace-dir parameter
2. Environment variable WORKSPACE_DIR
3. Environment variable PROJECT_ROOT
4. Environment variable CLAUDE_WORKSPACE
5. Current working directory

Articles are output by default to the {工作区}/articles/ directory.

Dependencies

• Python 3.9+
• requests

Install dependencies:

pip install requests

Output Summary

After the script completes, it outputs:

• Article title
• Author and official account information
• Output directory
• Markdown file path
• Number of downloaded images
• Formatting change summary

Notes

• This skill is an independent implementation and does not depend on other skills
• If the markdown-formatter skill exists in the current project, when using this skill you should automatically run markdown-formatter once more after the script generates Markdown
• Mainly oriented toward WeChat public article page structures; ordinary web pages are not guaranteed to be fully compatible
• If the article is restricted, deleted, or requires login, scraping will fail
• Markdown will be written directly to the target file in the output directory