A FastAPI service that reverse-engineers WeChat RSS

What it does

This FastAPI service logs into the WeChat public-platform backend via QR code, then scrapes articles from any official account and repackages them as standard RSS 2.0 feeds, Markdown exports, or JSON endpoints. It also proxies images, searches accounts by name, and polls for new content on a schedule. The whole thing is Dockerized and can run locally or on a server.

The interesting bit

The anti-ban stack is the standout: it rotates SOCKS5 proxies, simulates Chrome TLS fingerprints, and applies three layers of automatic rate limiting to keep WeChat’s risk controls at bay. That level of defensive scraping engineering is usually hidden in proprietary tools, not open-sourced AGPL code. The project also advertises MCP support for plugging into AI agents, though the README truncates before showing the mechanics.

Key highlights

• Full-content RSS feeds with images, compatible with FreshRSS, Feedly, and NetNewsWire
• Markdown export with YAML frontmatter and incremental sync for Obsidian or Logseq
• Chrome TLS fingerprint simulation plus SOCKS5 proxy rotation to evade WeChat rate limits
• Account search, article search within accounts, and metadata like verification status and original article count
• Webhook alerts when the mandatory QR-code login is about to expire (roughly every four days)

Caveats

• You must own a WeChat official account (subscription or service) just to log in and generate credentials
• The QR-code login session lasts only about four days, so you’ll need to re-scan periodically
• Remote access requires a public server or reverse proxy; local-only mode works on localhost

Verdict

Worth a look if you curate news outside the WeChat app or pipe content into a knowledge base. Skip it if you don’t have a WeChat official account to spare or if periodic QR-code re-authentication sounds like a chore.