LangChain 1.0 as an 8-Week Bootcamp, in Repo Form

What it does This repository is a semester-style learning path for LangChain 1.0 and LangGraph 1.0, organized into 22 modules and three capstone projects. It treats GitHub like a learning management system, walking through agent loops, RAG, memory, middleware, and multi-agent graphs in four phased stages. All materials are written in Chinese.

The interesting bit Rather than dumping disconnected snippets, the author mapped an eight-week curriculum onto directories, complete with week-by-week pacing. It also pulls double duty as a migration guide, explicitly warning that langgraph.prebuilt.create_react_agent is deprecated and routing readers to the new langchain.agents.create_agent API.

Key highlights

• 22 modules covering fundamentals through advanced topics (structured output, checkpointing, mixed modality, error handling)
• 3 capstone projects: a RAG document system, a multi-agent customer support bot, and a research assistant
• Explicit LangChain 1.0 API callouts, including middleware hooks (before_model, after_model) and the new create_agent entrypoint
• Fully localized for Chinese-speaking developers, from README to code comments

Caveats

• The README lists an October 2025 release date for LangChain 1.0, which may confuse readers
• Full-stack front-end and back-end projects (FastAPI, Vue, MCP) are listed as “follow-up plans,” not yet implemented
• Examples require several external API keys (Groq, OpenAI, Pinecone) to run

Verdict Worth bookmarking if you are a Chinese-speaking developer trying to get oriented on the new LangChain 1.0 stack quickly. Skip it if you need a reusable library—this is coursework, not code to import.