LangChain & LangGraph
Development Environment
A complete step-by-step guide to set up your local environment for building LLM-powered applications and agent workflows.
Python Version & Virtual Environment
PrerequisitesLangChain requires Python 3.9+. Use pyenv or your system package manager to install the correct version, then create an isolated virtual environment.
# Check Python version python --version # Python 3.11.x recommended # Create & activate a virtual environment python -m venv .venv source .venv/bin/activate # macOS / Linux # .venv\Scripts\activate # Windows # Upgrade pip first pip install --upgrade pip
Install LangChain Core Packages
InstallationInstall the main LangChain packages. langchain provides the high-level API, langchain-core is the base layer, and langchain-community bundles third-party integrations.
pip install langchain langchain-core langchain-community # Provider-specific packages (install what you need) pip install langchain-openai # OpenAI / Azure OpenAI pip install langchain-anthropic # Anthropic Claude pip install langchain-google-genai # Google Gemini pip install langchain-ollama # Local models via Ollama
Install LangGraph
Agents & FlowsLangGraph extends LangChain with a graph-based orchestration layer for building stateful, multi-actor agent workflows with cycles, branching, and persistence.
pip install langgraph # Optional: LangGraph CLI for local dev server pip install langgraph-cli # Optional: checkpoint persistence backends pip install langgraph-checkpoint-sqlite # SQLite (easy local) pip install langgraph-checkpoint-postgres # PostgreSQL (prod)
Configure API Keys
SecretsNever hard-code keys in source files. Use a .env file loaded with python-dotenv, or export them as shell environment variables.
# .env (add to .gitignore!) OPENAI_API_KEY=sk-... ANTHROPIC_API_KEY=sk-ant-... LANGCHAIN_API_KEY=ls__... # LangSmith tracing LANGCHAIN_TRACING_V2=true LANGCHAIN_PROJECT=my-project
pip install python-dotenv # At the top of every script / notebook from dotenv import load_dotenv load_dotenv() # reads .env into os.environ
Verify the Installation
Smoke TestRun this quick sanity check to confirm all packages are importable and the LLM responds correctly.
from dotenv import load_dotenv load_dotenv() from langchain_openai import ChatOpenAI from langchain_core.messages import HumanMessage llm = ChatOpenAI(model="gpt-4o-mini") response = llm.invoke([HumanMessage(content="Hello, LangChain!")]) print(response.content)
AuthenticationError, double-check that your .env file is in the project root and load_dotenv() runs before the model call.
Your First LangGraph Agent
Hello GraphA minimal LangGraph workflow: a single-node graph that routes a user message through an LLM and returns the response.
from typing import TypedDict, Annotated from langgraph.graph import StateGraph, END from langchain_openai import ChatOpenAI from langchain_core.messages import HumanMessage, BaseMessage import operator class State(TypedDict): messages: Annotated[list[BaseMessage], operator.add] llm = ChatOpenAI(model="gpt-4o-mini") def chat_node(state: State) -> State: response = llm.invoke(state["messages"]) return {"messages": [response]} graph = StateGraph(State) graph.add_node("chat", chat_node) graph.set_entry_point("chat") graph.add_edge("chat", END) app = graph.compile() result = app.invoke({"messages": [HumanMessage("What is LangGraph?")]}) print(result["messages"][-1].content)
Recommended Tools & Next Steps
EcosystemYour environment is ready. Here’s what to explore next:
- LangSmith — observability, tracing & prompt management at
smith.langchain.com - LangGraph Studio — visual graph editor & local dev server (
langgraph dev) - LangChain Hub — community prompt library at
hub.langchain.com - LCEL — LangChain Expression Language for composing chains with
|pipes - Vector Stores — add
langchain-chroma,langchain-pinecone, orfaiss-cpufor RAG - requirements.txt — freeze deps with
pip freeze > requirements.txtfor reproducibility
