Master the OpenAI API and Build Your First Intelligent Chatbot

How to Use the OpenAI API

The OpenAI API allows access to models like GPT-4, GPT-3.5, and DALL-E. Here’s how to get started:

Step 1: Get an API Key

Visit platform.openai.com and create an account.

In the "API Keys" menu, click "Create new secret key."

Copy the generated key and store it securely.

Step 2: Install the OpenAI Library

Run the following command in the terminal (with the virtual environment active):

pip install openai

Step 3: First API Call

Create a file chatbot.py and add:

import openai  

openai.api_key = "your-api-key"  

response = openai.chat.completions.create(  
    model="gpt-3.5-turbo",  
    messages=[{"role": "user", "content": "Who invented the light bulb?"}]  
)  

print(response.choices[0].message.content)

Expected Output:

"Thomas Edison is credited as the inventor of the practical incandescent light bulb in 1879."

Creating an Interactive Chatbot with GPT

Let’s transform the above code into an interactive chatbot.

Step 1: Chatbot in a Loop

Update chatbot.py:

import openai  

openai.api_key = "your-api-key"  

history = []  

while True:  
    question = input("\nYou: ")  
    if question.lower() == "exit":  
        break  

    history.append({"role": "user", "content": question})  

    response = openai.chat.completions.create(  
        model="gpt-3.5-turbo",  
        messages=history,  
        temperature=0.7  
    )  

    ai_response = response.choices[0].message.content  
    print(f"\nAI: {ai_response}")  
    history.append({"role": "assistant", "content": ai_response})

How It Works:

  • history stores the conversation context.

  • temperature=0.7 controls creativity (0 = conservative, 1 = random).

Example Usage:

You: How do I bake a chocolate cake?
AI: To bake a chocolate cake, you’ll need...

Customizing Responses

Adjust chatbot behavior using system prompts and advanced parameters.

Example 1: Sarcastic Assistant

Add a system prompt to the history:

history = [
    {"role": "system", "content": "You are a sarcastic assistant who responds with irony."}
]

Output:

You: What is the meaning of life?
AI: Oh sure, because I definitely have the answer to humanity’s biggest question...

Example 2: Controlling Style and Format

Request responses in specific topics or styles:

question = "Explain relativity theory in 3 points, like a pirate."

Output:

1. Arr! General relativity says gravity ain't no force, but space-time curving...

Advanced Parameters:

  • max_tokens=150: Limits response size.

  • top_p=0.9: Controls word choice diversity.

Practical Project: Study Assistant

Create an AI Agent that simplifies complex concepts:

history = [
    {"role": "system", "content": "You are a science tutor for 10-year-olds. Use simple analogies."}
]

question = "What is a black hole?"

Output:

"Imagine a giant vacuum cleaner in space that even light can't escape! It’s like an invisible monster swallowing everything."

Common Errors and Fixes

Authentication Error:

Error:

openai.AuthenticationError: Invalid API key

Solution: Ensure the API key is correct and active.

Token Limit Exceeded:

Error:

openai.BadRequestError: maximum context length is 4097 tokens

Solution: Reduce history size or use max_tokens.

Rate Limit Exceeded:

Error:

openai.RateLimitError

Solution: Wait 20 seconds between requests or upgrade to a paid plan.

Next Steps

In the next chapter, you’ll learn to add memory to your AI Agent using databases. Meanwhile:

✅ Try creating a themed bot (e.g., poet, financial advisor). ✅ Explore OpenAI documentation for more parameters.

Pro Tips:

  • Use pip install python-dotenv to store your API key in a .env file and keep it secure.

  • System prompts define your AI Agent’s personality.

  • The temperature parameter adjusts creativity.

  • Always manage conversation history to maintain context.

Ready to upgrade your chatbot? In the next chapter, we’ll transform it into an agent with memory!

 

Copyright Statement: Unless stated otherwise, all articles are original to this site, please credit the source when sharing.

Article link:http://pybeginners.com/artificial-intelligence/creating-a-simple-ai-agent-with-openai-api/

License Agreement:Attribution-NonCommercial 4.0 International License