Getting Started with ChatGPT

Getting Started with ChatGPT

Go to the ChatGPT website, create an account, and open a conversation. The free tier includes a capable model, while paid plans add faster responses, priority access, and higher usage limits. Start with short, clear questions and read the answers critically.

Set Up Your Workspace

• Use multiple chats per topic so each conversation keeps a clean context.
• Pin important chats and name them for easy search later.
• Use the share feature to send a conversation to a teammate or as a bug report.

Choose the Right Style for the Job

The model follows the role and format you describe. If you ask for a step list, you get a list. If you ask for a JSON schema, you get JSON. Explicit instructions in the message shape the output more than any hidden setting.

from openai import OpenAI
client = OpenAI()
completion = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {"role": "system", "content": "Answer in three bullet points."},
        {"role": "user", "content": "What is a SQL index?"}
    ]
)
print(completion.choices[0].message.content)

A Simple Workflow

Describe the goal, paste relevant code, ask for a plan, and then ask follow-up questions instead of starting over. Keep a conversation focused so the model keeps the important details in context. When a chat grows long and answers drift, start a new one and restate the goal and the constraints.

Learn the Interface

You can edit your own message after sending, which is useful when you make a typo. Use the regenerate button when you suspect the first answer was weak; the model will sample a new output. Copy code blocks directly rather than retyping them to avoid transcription errors.

Key Points

• Create an account and keep conversations focused per topic.
• State the desired role, format, and constraints in your message.
• Edit and regenerate messages to steer answers.
• Verify everything before you trust it in production.