WeChat bot plumbing for polyglot AI agents

What it does

This SDK family wraps WeChat’s iLink Bot protocol so your code can read and write messages, files, and voice notes without babysitting QR-code logins, HTTP long polling, or session expiry. It is pitched at AI agents—specifically OpenClaw-style workflows—but any automated bot fits. The Node.js version carries the most bells and whistles, while Python, Go, and Rust aim for parity on the core client.

The interesting bit

The value is in the boring stuff: automatic cursor management for long polling, AES-128-ECB CDN encryption, and retry logic when WeChat throws the dreaded -14 session error. That is exactly the kind of low-level protocol bookkeeping that usually rots in a private gist, not a versioned SDK. The Node.js port even layers on Express-style middleware and a chainable message builder, implying the authors expect you to build real products on top of it.

Key highlights

• Multi-language core: TypeScript/Node.js, Python, Go, and Rust.
• Stateful protocol hygiene: QR login, credential persistence, context token lifecycle, and automatic session recovery.
• Rich media: upload and download images, files, voice, and video.
• Node.js extras: middleware pipeline, pluggable storage, structured logging, and typed events.
• Pi Agent extension: a drop-in bridge to the Pi programming assistant.

Caveats

• The README does not spell out the relationship between the iLink protocol and Tencent’s official WeChat API program, which matters for anyone evaluating vendor risk.
• Documentation is overwhelmingly Chinese; the English README is linked but not surfaced in the main source.
• All four SDKs are tagged “production-ready,” though the source only lists test file counts without coverage or stability metrics.

Verdict

Pick this up if you need an AI agent inside WeChat and you are tired of maintaining protocol glue. Look elsewhere if you need a vendor-supported WeChat Business API or comprehensive English docs.