ai-agents-from-zero
A continuously updated Chinese-language tutorial series that walks through building AI agents from the ground up to enterprise deployment, covering LangChain, LangGraph, Coze, Dify, MCP and RAG.
This is a long-term, openly maintained course pairing lessons with runnable source code, an interview question bank, and enterprise-scale example projects, aimed at an 'AI agent / large-model application developer' role. As of this material, its concepts unit is complete, along with two full example projects: a natural-language-to-SQL e-commerce project built with LangGraph, and a multi-agent deep-research project built with DeepAgents.
Its technical path runs from large models and prompt engineering through low-code platforms such as Coze and Dify and code frameworks such as LangChain, LangGraph and DeepAgents, to enterprise RAG and agent design, fine-tuning and engineering practices, focused on a Python stack rather than Java frameworks such as Spring AI or langchain4j. The example projects also cover intent parsing, multi-source knowledge retrieval, escalation to a human, and monitoring.
It is written in Python and licensed under MIT.