Weixin

This MCP server provides a browser-based spider that collects and analyzes WeChat public account articles, enabling AI agents to access dynamic content through a standard MCP interface. It automates article crawling, image handling, and content analysis so you can build workflows that summarize, index, or extract insights from WeChat content.

How to use

Start the MCP server and connect with your MCP client to begin crawling WeChat public account articles. The server exposes three core capabilities: Article Crawling to fetch articles, Content Analysis to extract keywords and summaries, and Statistics to provide counts and trends. Use your MCP client to request crawling tasks, then retrieve processed results, including downloaded images stored locally.

How to install

Prerequisites: Python 3.8+ (3.10+ recommended) and a Chrome/Chromium browser for the automated spider.

Step 1: Install Python dependencies.

pip install -r requirements.txt

Step 2: Prepare and run the MCP server module.

python -m mcp_weixin_spider

Step 3: If you want to run the client demo separately, execute this in another terminal.

python -m mcp_weixin_spider.client

Additional sections

Configuration shown below defines how the MCP server starts and which environment variables it uses. Place this in your MCP configuration. The server is started as a stdio-based MCP endpoint using a Python command and a path to the spider script.

{
  "mcpServers": {
    "weixin_spider": {
      "command": "python",
      "args": [
        "本地路径/server.py"
      ],
      "env": {
        "ARTICLES_DIR": "articles",
        "DOWNLOAD_IMAGES": "true",
        "HEADLESS": "true",
        "WAIT_TIME": "10"
      }
    }
  }
}

Available tools

ArticleCrawler

Core tool that fetches and saves articles from WeChat public accounts, handling dynamic content via the Selenium spider.

ContentAnalyzer

Tool that analyzes article text to extract keywords, summaries, and other metadata.

StatisticsCollector

Tool that aggregates counts and statistics such as article counts, image counts, and word frequencies.