Calling an AI API with Python

Relax. We'll talk through this in plain words — no textbook voice.

What you'll walk away with

Understand Calling an AI API with Python without the intimidation factor
Get hands-on practice trying it yourself
Spot the common pitfalls and sidestep them with a smile

Build a command-line AI helper by calling the Responses API with the OpenAI Python SDK.

Let's think about it this way for a moment

The Python SDK follows the same flow as JavaScript — build a client, send the model and input, read output_text. Using a virtual environment keeps each project's dependencies from tangling with one another.

Calling an AI API with Python lesson illustration — AI APIs — Calling an AI API with Python

Let's connect this to everyday life

Create an environment with python -m venv .venv, run pip install openai, and keep your key in an environment variable. Never hard-code a secret into your code.

Create a virtual environment
Install the openai package
Set the environment variable
Catch response errors

Let's try it hands-on together

python

from openai import OpenAI

client = OpenAI()

response = client.responses.create(
    model="gpt-5.6",
    input="Git branch ကို beginner နားလည်အောင် မြန်မာလိုရှင်းပြပါ။",
)

print(response.output_text)

You should see

A Python script will print the AI's answer to the terminal.

5-minute try-it

Take a user's question with input() and call the API with it. Add validation and an exception message so an empty input never gets sent.

A quick word of caution

Don't treat AI output as the final, guaranteed-correct answer. For anything important — critical facts, code, or user data — have a human review it before you use it.

OpenAI API Quickstart — OpenAI

OpenAI — Using GPT-5.6 — OpenAI

Easy traps

Installing every package into the global Python environment
Logging an API key or other sensitive data inside an exception message

Exercise

Take a user's question with input() and call the API with it. Add validation and an exception message so an empty input never gets sent.

You'll know it worked when: A Python script will print the AI's answer to the terminal.