Skip to content

ChatGPT

What it is

ChatGPT is an AI-powered chatbot developed by OpenAI. It uses large language models to provide human-like responses to various prompts and questions.

What problem it solves

Provides a general-purpose conversational AI interface for brainstorming, drafting text, answering questions, and completing a wide range of language tasks without requiring technical setup.

Where it fits in the stack

AI & Knowledge — used as a cloud-based conversational assistant alongside local LLM options.

Typical use cases

  • Drafting and editing text, emails, and documentation.
  • Answering research questions and summarizing information.
  • Excel Data Integration: Direct data processing and financial integrations within Excel (launched 2026-03-06).
  • Advanced Education: New ways to learn math and science with interactive step-by-step guidance (launched 2026-03-11).
  • Exploring Custom GPTs for repository-specific or domain-specific knowledge.

Strengths

  • Broad general knowledge and strong reasoning capabilities
  • Easy to use with no local infrastructure required
  • Extensible through Custom GPTs and plugin ecosystem

Limitations

  • Requires an internet connection and sends data to OpenAI servers
  • Not privacy-first; unsuitable for sensitive or private data without additional safeguards
  • Usage beyond the free tier requires a paid subscription or API credits

When to use it

  • When you need quick, high-quality text generation or brainstorming and privacy is not a concern
  • When local LLM infrastructure is unavailable or insufficient for the task

When not to use it

  • When working with sensitive or private data that should not leave the local network
  • When deterministic, reproducible outputs are required

Licensing and cost

  • Open Source: No (Proprietary).
  • Cost: Free tier available; "Plus" subscription for priority access and GPT-4o; usage-based API.
  • Self-hostable: No.

Getting started

Web Interface

  1. Visit chatgpt.com.
  2. Log in with an OpenAI, Google, or Microsoft account.
  3. Type a message in the chat bar to begin.

OpenAI API

For programmatic access to ChatGPT models (gpt-4o, gpt-4-turbo, etc.): 1. Create an account at platform.openai.com. 2. Add credits to your billing account. 3. Create an API key. 4. Install the library: pip install openai.

CLI examples

Official OpenAI CLI

OpenAI provides a CLI tool as part of the openai Python package:

# Set your API key
export OPENAI_API_KEY='your-api-key'

# Simple chat completion from terminal
openai api chat_completions.create -m gpt-4o -g user "What is the capital of France?"

# List available models
openai api models.list

Unofficial CLI (using shell-gpt)

A popular way to use ChatGPT for terminal automation:

# Install Shell GPT
pip install shell-gpt

# Use it as a filter (REPL style)
ls | sgpt "summarize these files"

# Request a shell command (Dangerous, use with caution)
sgpt --shell "find all python files larger than 1MB"

API examples

Python (OpenAI SDK)

Standard synchronous request for a chat completion:

from openai import OpenAI

client = OpenAI()

response = client.chat.completions.create(
    model="gpt-4o",
    messages=[
        {"role": "system", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Write a one-sentence summary of the Pydantic v2 changes."}
    ]
)

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

JSON Mode

Ensuring the model returns a valid JSON object for automation:

from openai import OpenAI

client = OpenAI()

response = client.chat.completions.create(
    model="gpt-4o",
    response_format={ "type": "json_object" },
    messages=[
        {"role": "system", "content": "You are a helpful assistant that outputs JSON."},
        {"role": "user", "content": "List the top 3 cities in Japan by population as JSON."}
    ]
)

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

Sources / references

Contribution Metadata

  • Last reviewed: 2026-05-20
  • Confidence: high