LangChain & LangGraph

LangChain has become the most popular framework for building LLM applications, with over 120,000 GitHub stars and a massive community. LangGraph extends it with stateful, multi-actor workflows.

LangChain: The Foundation

LangChain provides building blocks for LLM applications:

from langchain_openai import ChatOpenAI
from langchain.agents import create_tool_calling_agent, AgentExecutor
from langchain_core.prompts import ChatPromptTemplate

# Initialize the model
llm = ChatOpenAI(model="gpt-4")

# Define tools
tools = [search_tool, calculator_tool]

# Create the agent
prompt = ChatPromptTemplate.from_messages([
    ("system", "You are a helpful research assistant."),
    ("human", "{input}"),
    ("placeholder", "{agent_scratchpad}")
])

agent = create_tool_calling_agent(llm, tools, prompt)
executor = AgentExecutor(agent=agent, tools=tools)

# Run the agent
result = executor.invoke({"input": "What's the population of Tokyo?"})

Key LangChain Concepts

Concept	Description	Use Case
Chains	Sequences of operations	Linear workflows
Agents	Dynamic decision-makers	Complex tasks
Tools	External capabilities	Search, compute, APIs
Memory	Conversation history	Chatbots, assistants
Retrievers	Document fetching	RAG applications

LangGraph: Stateful Workflows

LangGraph adds graph-based state management for complex agent systems:

from langgraph.graph import StateGraph, END
from typing import TypedDict

class AgentState(TypedDict):
    messages: list
    current_step: str

# Define the graph
workflow = StateGraph(AgentState)

# Add nodes (processing steps)
workflow.add_node("research", research_node)
workflow.add_node("analyze", analyze_node)
workflow.add_node("write", write_node)

# Add edges (transitions)
workflow.add_edge("research", "analyze")
workflow.add_conditional_edges(
    "analyze",
    should_continue,
    {"continue": "write", "research_more": "research"}
)
workflow.add_edge("write", END)

# Compile and run
app = workflow.compile()
result = app.invoke({"messages": [], "current_step": "start"})

When to Use LangChain/LangGraph

✅ Good fit:

• Building production LLM applications
• Need extensive tool integrations
• Want a large community and ecosystem
• Require complex state management

⚠️ Consider alternatives when:

• You need minimal dependencies
• Building simple, one-off scripts
• Team unfamiliar with the abstraction style

LangSmith: Observability

LangSmith provides tracing and debugging:

# Enable tracing
import os
os.environ["LANGCHAIN_TRACING_V2"] = "true"
os.environ["LANGCHAIN_PROJECT"] = "my-agent-project"

# All agent runs are now traced automatically

Next, we'll explore CrewAI's role-based approach to multi-agent systems. :::