🦜🔗 LangChain LangGraph Crash Course

LangChain LangGraph Crash Course (LLCC) is a hands-on course designed to teach you LangChain and LangGraph from the ground up. In this course, you will learn how to build Retrieval-Augmented Generation (RAG) workflows, intelligent agents doing web searches, text classifiers, and much more.

Note

Unlike official tutorials, LLCC provides pure Python scripts instead of Jupyter or Colab notebooks, delivering clean, ready-to-run code. Each script is extensively commented for complete clarity and ease of understanding. Additionally, LLCC sometimes includes extra steps not found in official tutorials, such as LCEL.

Table of contents

• Requirements
• Install
• Cost
• Course
• Resources

Requirements

If you want to run the examples, you will need to install the following:

• Pixi

You'll also need to have an OpenAI key.

Install

Inside the repository, install the dependencies as follows:

pixi install -a

Cost

All examples utilize the OpenAI API, specifically employing the gpt-4o-mini model as the language model and/or the text-embedding-3-large model for embeddings. You can access these services under the free tier, which typically incurs no cost.

Course

Tip

It's recommended to follow the course in the given order.

1. LangChain

Task	Description	File
Chat models & prompts	Simple LLM application with prompt templates and chat models	simple_llm_application.py
Semantic search	Search over a PDF with document loaders, embedding models and vector stores	semantic_search_engine.py
Classification	Classify text into tags using chat models with structured outputs	text_classifier.py
Extraction	Extract structured data from text using chat models and few-shot examples	data_extractor.py

2. LangGraph

2.1 Agents

Task	Description	File
Agent	Simple agent with a memory and able to do web searches	agent.py
Agent & human in the loop	Agent empowered to request a human intervention	agent_human_assitance.py
Agent & time travel	Altering an agent output by changing its memory	agent_time_travel.py

2.2 Retrieval augmented generation (RAG)

Task	Description	File
RAG	Simple RAG with an introduction to self-query (an advanced RAG technique)	rag.py
RAG & conversations	Delegating (multi-step) RAG calls to a LLM	rag_delegation.py

Resources

The resources described here are sorted in order of importance. For instance, to better understand and grasp each LangChain concept, you should have knowledge of LLM theory and LLM engineering.

LLM theory

• LLM course

LLM engineering

• LLM's Engineer Handbook

LangChain

• LangChain tutorials
• LangChain expression language (LCEL)

LangGraph

• LangGraph tutorials