Build a Tool-Calling AI Agent with LangChain, LangGraph and OpenAI Models

📘 Introduction

Welcome to this hands-on LangChain + LangGraph tutorial! In this guide, we’ll build an AI agent with tool-calling capabilities using LangChain, LangGraph, and OpenAI models.

Instead of just answering questions, our agent will be able to invoke a Python function when needed. We’ll demonstrate this with a simple tool that greets users — the perfect starting point to understand how tool integration works.

💡 What are LangChain and LangGraph?

• LangChain is a Python framework for building applications powered by LLMs. It gives you tools to define prompts, memory, tools, and workflows — making it possible to build more than just chatbots.
• LangGraph extends LangChain by allowing graph-based workflows. You can define nodes (steps) and edges(transitions) to control how agents reason, plan, and act.

Together with OpenAI models, this stack allows you to:

🤖 Build context-aware agents
🛠️ Equip them with tools (e.g., APIs, calculators, or custom Python functions)
🔗 Control workflows with graph structures
⚡ Scale from simple prototypes to powerful assistants

✅ Prerequisites

Before starting, make sure you have the following:

🐍☑️ Python installed
📦☑️ Pip installed
🌐☑️ A virtual environment (venv) created and activated
🔑☑️ An OpenAI API key generated (get one from OpenAI)
🔐☑️ OpenAI API key set as environment variable OPENAI_API_KEY

📦1️⃣ Install Libraries

Install the following Python packages using pip:

pip install langchain-openai langgraph

📥2️⃣ Import Libraries

Start by importing the required Python modules:

from langchain_openai import ChatOpenAI
from langgraph.prebuilt import create_react_agent
from langchain_core.tools import tool

🛠️3️⃣ Define a Tool

We’ll create a simple Python greeting function and register it as a tool so that the agent can call it when needed.

@tool
def greet_user(name: str) -> str:
    """Greet the user with a personalized welcome message.

    Args:
        name: The name of the user.
    """
    return f"Hello, {name}! Welcome to the Deep Learning Nerds Academy"