🧠 Building Your Own ChatGPT Clone in Python using OpenAI API

🚀 Introduction

Have you ever wondered how ChatGPT actually works under the hood?
What if you could build your own ChatGPT-like chatbot — fully customizable, interactive, and smart enough to carry out conversations in real-time?

In this guide, we’ll walk you through building a ChatGPT clone using Python and the OpenAI API.
By the end, you’ll have your own interactive chatbot that can:

Accept custom system prompts (change the assistant’s behavior)
Maintain chat history
Switch personality mid-conversation
Exit gracefully when you’re done

Let’s get started! 💬

🧩 Prerequisites

Before you start, make sure you have:

🐍 Python 3.8+
🔑 An OpenAI API key (you can get it from OpenAI Dashboard)
📦 Required Python libraries:
```
pip install openai streamlit
```

💡 Project Overview

We’ll build a terminal-based ChatGPT clone that uses OpenAI’s gpt-3.5-turbo model.
This app will:

Accept a system prompt to define the assistant’s behavior.
Allow you to chat interactively.
Let you update the system prompt dynamically.
Gracefully exit on “bye”, “exit”, or “quit”.

Later, you can easily adapt this to a Streamlit-based web interface.

🧠 Step 1: Import Dependencies

Let’s begin by importing the required libraries:


import streamlit as st
from openai import OpenAI
import os
import time

openai → To access GPT models
os → To handle environment variables (for API keys)
time → To add small response delays
streamlit → Optional, for building a web interface later

🤖 Step 2: Define a Simple Chat Function

This function sends a single prompt to the GPT model and returns the response:


def gpt_chat(prompt):
    client = OpenAI(api_key=os.environ.get('OPENAI_API_KEY'))
    messages = list()
    system_prompt = "Answer as concisely as possible"
    messages.append({"role": "system", "content": system_prompt})
    messages.append({"role": "user", "content": prompt})
    
    response = client.chat.completions.create(
        model="gpt-3.5-turbo",
        messages=messages,
        temperature=0.7,
        max_tokens=300
    )
    current_response = response.choices[0].message.content
    return current_response

💡 Tip:
You can modify the system_prompt to change the chatbot’s personality —
for example, “You are a sarcastic assistant” or “You are a helpful teacher”.

💬 Step 3: Building the Interactive Chat Loop

This is where the magic happens — we maintain a message history and allow users to chat freely.


def interactive_chat():
    client = OpenAI(api_key=os.environ.get('OPENAI_API_KEY'))
    questions = list()
    bot_responses = list()
    messages = list()
    
    # Get system prompt from user
    print("Enter system prompt (or press Enter for default):")
    try:
        system_prompt = input("System: ").strip()
        if not system_prompt:
            system_prompt = "Answer as concisely as possible"
        print(f"Using system prompt: '{system_prompt}'")
    except EOFError:
        system_prompt = "Answer as concisely as possible"
        print(f"Using default system prompt: '{system_prompt}'")
    
    messages.append({"role": "system", "content": system_prompt})
    print("-" * 50)
    
    try:
        while True:
            current_question = input("Me: ")
            if current_question.lower() in ["exit", "quit", "bye"]:
                print("Bot: Bye!")
                time.sleep(2)
                break

            # Allow changing system prompt mid-chat
            if current_question.lower() in ["/system", "/prompt"]:
                print("Enter new system prompt:")
                try:
                    new_prompt = input("System: ").strip()
                    if new_prompt:
                        system_prompt = new_prompt
                        messages[0] = {"role": "system", "content": system_prompt}
                        print(f"System prompt updated to: '{system_prompt}'")
                    else:
                        print("System prompt unchanged.")
                except EOFError:
                    print("System prompt unchanged.")
                print("-" * 50)
                continue
            
            if current_question == '':
                continue

            messages.append({"role": "user", "content": current_question})
            response = client.chat.completions.create(
                model="gpt-3.5-turbo",
                messages=messages,
                temperature=0.7,
                max_tokens=300
            )
            current_response = response.choices[0].message.content
            print(f"Bot: {current_response}")
            bot_responses.append(current_response)
            messages.append({"role": "assistant", "content": current_response})
            questions.append(current_question)
            print('\n' + '-' * 50 + '\n')
    except EOFError:
        print("\nInteractive chat ended (EOF detected).")
    except KeyboardInterrupt:
        print("\n\nChat ended by user (Ctrl+C).")

✨ Highlights:

/system or /prompt allows you to update the system prompt anytime.
Conversation context is preserved with messages history.
Clean exit with bye, exit, or quit.

🔐 Step 4: Running the Chat

To kickstart your chatbot, add the following:


if __name__ == "__main__":
    # Set API key
    os.environ['OPENAI_API_KEY'] = 'your-openai-api-key-here'
    print("Welcome to ChatGPT Clone!")
    print("Features:")
    print("- Set custom system prompt at startup")
    print("- Change system prompt during chat with '/system' or '/prompt'")
    print("- Type 'exit', 'quit', or 'bye' to end the conversation")
    print("- Press Ctrl+C for emergency exit")
    print("=" * 50)
    interactive_chat()

Run it in your terminal:


python chatgpt_clone.py

🧭 Example Conversation


Welcome to ChatGPT Clone!
Features:
- Set custom system prompt at startup
- Change system prompt during chat with '/system' or '/prompt'
==================================================

Enter system prompt (or press Enter for default):
System: You are a witty assistant who gives humorous answers
Using system prompt: 'You are a witty assistant who gives humorous answers'
--------------------------------------------------
Me: What's the weather today?
Bot: I’m not a weatherman, but it feels like “stay indoors with coffee” kind of day! ☕
--------------------------------------------------
Me: /system
Enter new system prompt:
System: You are a professional software mentor
System prompt updated to: 'You are a professional software mentor'
--------------------------------------------------
Me: Explain Python decorators
Bot: Decorators in Python are functions that modify the behavior of another function without changing its code. They’re often used for logging, access control, or measuring execution time.
--------------------------------------------------
Me: bye
Bot: Bye!

🌐 Bonus: Streamlit Web Interface (Optional)

You can easily turn this into a web-based chatbot using Streamlit.
Here’s a mini version:


import streamlit as st

st.title("💬 ChatGPT Clone")

if "messages" not in st.session_state:
    st.session_state.messages = [{"role": "system", "content": "Answer concisely"}]

user_input = st.text_input("You:")

if user_input:
    st.session_state.messages.append({"role": "user", "content": user_input})
    response = gpt_chat(user_input)
    st.session_state.messages.append({"role": "assistant", "content": response})
    st.write(f"**Bot:** {response}")

Run it with:


streamlit run app.py

🏁 Conclusion

🎉 Congratulations! You’ve just built your own ChatGPT clone using Python and OpenAI’s API.
You now understand:

How to manage chat history
How to use system prompts to alter behavior
How to extend the terminal chatbot into a web app

This setup forms the backbone of AI-driven assistants, customer support bots, and educational tutors — all powered by GPT.

Table of Contents