How to beginner · 3 min read

How to create an agent in LangChain

Q: How to create an agent in LangChain

To create an agent in LangChain, instantiate a language model like ChatOpenAI and combine it with tools and an agent class such as initialize_agent. This enables dynamic decision-making and tool usage within your AI workflows.

Quick answer

To create an agent in LangChain, instantiate a language model like ChatOpenAI and combine it with tools and an agent class such as initialize_agent. This enables dynamic decision-making and tool usage within your AI workflows.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install langchain_openai langchain_community openai

Setup

Install the necessary packages and set your OpenAI API key as an environment variable.

Install LangChain and OpenAI SDK: pip install langchain_openai langchain_community openai
Set environment variable in your shell: export OPENAI_API_KEY='your_api_key' (Linux/macOS) or setx OPENAI_API_KEY "your_api_key" (Windows)

bash

pip install langchain_openai langchain_community openai

Step by step

This example creates a simple agent that uses the OpenAI gpt-4o model and a calculator tool to answer questions and perform calculations.

python

import os
from langchain_openai import ChatOpenAI
from langchain_community.agents import initialize_agent
from langchain_community.tools import Calculator

# Initialize the language model
llm = ChatOpenAI(model_name="gpt-4o", temperature=0, openai_api_key=os.environ["OPENAI_API_KEY"])

# Initialize tools
calculator = Calculator()

# Create an agent with the tools and the language model
agent = initialize_agent(
    tools=[calculator],
    llm=llm,
    agent_type="zero-shot-react-description",
    verbose=True
)

# Run the agent with a query
response = agent.run("What is 25 multiplied by 4?")
print(response)

output

Common variations

You can customize agents by:

Using different models like gpt-4o-mini or gemini-1.5-pro.
Adding multiple tools such as search, database queries, or custom APIs.
Using async calls with asyncio for concurrency.
Changing the agent type to chat-zero-shot-react-description or conversational-react-description for different reasoning styles.

python

import asyncio

async def async_agent_example():
    llm = ChatOpenAI(model_name="gpt-4o-mini", temperature=0, openai_api_key=os.environ["OPENAI_API_KEY"])
    calculator = Calculator()
    agent = initialize_agent(
        tools=[calculator],
        llm=llm,
        agent_type="zero-shot-react-description",
        verbose=False
    )
    response = await agent.arun("Calculate 12 divided by 3.")
    print(response)

asyncio.run(async_agent_example())

output

4.0

Troubleshooting

If you get an authentication error, verify your OPENAI_API_KEY environment variable is set correctly.
If the agent returns unexpected answers, try lowering the temperature or adding more descriptive tool documentation.
For timeout errors, check your network connection and consider increasing request timeouts in the client.

✅

Key Takeaways

Use initialize_agent with ChatOpenAI and tools to create powerful LangChain agents.
Customize agents by changing models, tools, and agent types for your use case.
Always set your API key in os.environ to avoid authentication issues.

Verified 2026-04 · gpt-4o, gpt-4o-mini, gemini-1.5-pro

Verify ↗