Based on the official courses of LangChain Academy, each Jupyter Notebook is analyzed one by one, and the code logic and design principles are truly explained in Chinese - not translation, but analysis.
Each Module Summary has a built-in end-to-end practical project, which connects all the knowledge points of the Module into a complete, directly runnable real system.
A travel planning assistant capable of multi-step reasoning, automatic tool calling, and cross-wheel memory of user preferences. All six core knowledge points of Module 1 are connected into a complete runnable project, covering the complete link from StateGraph construction to Studio deployment.
An intelligent assistant that supports multiple rounds of dialogue, automatic history compression, and cross-session persistent memory. Organically integrating all six core knowledge points of Module 2, it is a LangGraph project skeleton that can be directly used in the production environment.
A contract processing workflow that integrates streaming progress display, manual approval, status editing, dynamic risk interruption and historical review. Integrate all 5 core knowledge points of Module 3 to simulate real enterprise-level approval scenarios.
A learning assistant that can continue to understand you across multiple conversations - record knowledge background (Profile), accumulate study notes (Collection), and let LLM decide independently when to extract and retrieve memories (Memory Agent).
It's not a translation, it's an in-depth analysis - the design intent and code principles behind each Notebook will be completely dismantled.
First establish awareness of the basic components of Agent, and then go deep into schema orchestration and production deployment. The two paths are connected one after the other, and both are indispensable.
Establish the basic capabilities of Agent application: model calling, prompt engineering, tool usage, MCP protocol, RAG knowledge retrieval, SQL query, middleware and manual approval process.
Continue to learn graph orchestration based on LangChain: StateGraph, state management, Human-in-the-Loop intervention, parallel execution, cross-session long-term memory and production environment deployment.
Click on any module card to jump directly and quickly locate the content you need.
Suggested learning paths:LangChain is responsible for understanding the basic components of Agent - model invocation, tool usage, memory management, and multi-Agent collaboration. On this basis, LangGraph teaches you to organize these components into complex workflows that are controllable, recoverable, and can run stably in a production environment.
Covering the complete AI Agent development skills stack from basic components to production deployment, the green markers are from the LangChain path and the purple markers are from the LangGraph path.