Agent School: A Chinese Curriculum for LangChain and LangGraph

What it does

Easy-langent is a Chinese-language tutorial—the name mashes “lang” and “agent” together—for developers who already know Python and basic LLM concepts but struggle to turn that into working agent code. It spans eight chapters, moving from LangChain fundamentals and RAG workflows to stateful LangGraph patterns and multi-agent collaboration, with each section tied to hands-on tasks rather than theory lectures.

The interesting bit

The curriculum treats social-deduction games as a teaching scaffold: learners build a “Who’s the Spy” agent and can study peer projects that include werewolf moderators, four-player script-murder games, and debate bots. Using adversarial multi-agent scenarios to teach state management and tool use is a genuinely clever pedagogical hack.

Key highlights

• Eight-chapter progression from framework basics to multi-agent systems
• Every chapter includes targeted build tasks instead of passive reading
• Peer project gallery spans RAG, medical diagnosis, customer-service tickets, and social-deduction games
• Maintained by Datawhale, a Chinese open-source education collective
• Hosted on GitHub Pages with a domestic China mirror for accessibility

Caveats

• Content is entirely in Chinese with no English translation visible
• Licensed under CC BY-NC-SA 4.0, so explicitly non-commercial
• Requires prior Python and basic LLM/agent knowledge; total beginners must finish prerequisite courses first

Verdict

Worth bookmarking if you read Chinese and want a structured, project-heavy path from LangChain confusion to shipping multi-agent workflows. If you need a drop-in framework or English documentation, look elsewhere.