How to beginner · 3 min read

How to define a tool for OpenAI API

Quick answer

To define a tool for the OpenAI API, create a Python function that wraps the OpenAI client calls using the SDK v1 pattern. Initialize the client with OpenAI(api_key=os.environ["OPENAI_API_KEY"]) and call client.chat.completions.create() with your desired model and messages.

PREREQUISITES

Python 3.8+
OpenAI API key (free tier works)
pip install openai>=1.0

Setup

Install the official OpenAI Python SDK version 1 or higher and set your API key as an environment variable.

Run pip install openai>=1.0 to install the SDK.
Set your API key in your environment: export OPENAI_API_KEY='your_api_key_here' (Linux/macOS) or setx OPENAI_API_KEY "your_api_key_here" (Windows).

bash

pip install openai>=1.0

Step by step

Define a Python function that acts as a tool to send prompts to the OpenAI chat completion endpoint using the gpt-4o model. This function initializes the client, sends the request, and returns the response text.

python

import os
from openai import OpenAI

# Define a tool function to call OpenAI chat completions

def openai_chat_tool(prompt: str) -> str:
    client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
    response = client.chat.completions.create(
        model="gpt-4o",
        messages=[{"role": "user", "content": prompt}]
    )
    return response.choices[0].message.content


if __name__ == "__main__":
    user_prompt = "Explain how to define a tool for OpenAI API in Python."
    result = openai_chat_tool(user_prompt)
    print("Response from OpenAI:")
    print(result)

output

Response from OpenAI:
To define a tool for the OpenAI API in Python, create a function that initializes the OpenAI client with your API key, sends a chat completion request using the desired model and prompt, and returns the generated response text.

Common variations

You can customize your tool by using different models like gpt-4o-mini, adding parameters such as max_tokens or temperature, or implementing asynchronous calls with asyncio. Streaming responses is also supported via the SDK's streaming interface.

python

import os
import asyncio
from openai import OpenAI

async def openai_chat_tool_async(prompt: str) -> str:
    client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
    response = await client.chat.completions.acreate(
        model="gpt-4o-mini",
        messages=[{"role": "user", "content": prompt}],
        max_tokens=150,
        temperature=0.7
    )
    return response.choices[0].message.content


if __name__ == "__main__":
    prompt = "Write a short poem about AI tools."
    result = asyncio.run(openai_chat_tool_async(prompt))
    print("Async response:")
    print(result)

output

Async response:
AI tools craft words with care,
Bringing thoughts from thin air.
Code and chat, they intertwine,
Making tasks swift and fine.

Troubleshooting

If you get an authentication error, verify your OPENAI_API_KEY environment variable is set correctly.
For rate limit errors, consider adding retry logic or reducing request frequency.
If the model name is invalid, confirm you are using a current model like gpt-4o.

✅

Key Takeaways

Use the official OpenAI Python SDK v1 with OpenAI client and environment variable API keys.
Wrap client.chat.completions.create() calls in a Python function to define your tool.
Customize your tool with different models, parameters, and async support as needed.

Verified 2026-04 · gpt-4o, gpt-4o-mini

Verify ↗