What are the CLI, Git and GitHub? Three different tools you will keep seeing in AI projects

AI tutorials often jump straight into instructions such as “open a terminal, clone the repository and checkout the branch.”

If those words are new, it can feel as though three unrelated skills are being taught at once.

They are related, but CLI, Git and GitHub are different things.

CLI: talk to the computer with commands

A Command-Line Interface (CLI) is an interface where you type commands instead of clicking buttons.

On Windows you may use PowerShell or Command Prompt. On macOS and Linux you commonly use a terminal shell.

A command can change folders, run Python, install a package, start a server or execute a development tool.

For example, “python app.py” means: ask the Python program to run the file app.py.

The important beginner habit is to know which folder you are currently in. Many “file not found” errors are simply caused by running a command from the wrong directory.

Git: version control on your project

Git is a distributed version-control system.

It records the history of files so you can see what changed, create a checkpoint and move between lines of work.

Three beginner concepts matter immediately:

• repository (repo) — the project tracked by Git,
• commit — a recorded snapshot/change with a message,
• branch — a named line of development.

A commit is not the same as saving a file. You can save many edits locally before deciding which changes belong in one commit.

GitHub: an online service built around Git repositories

GitHub hosts Git repositories and adds collaboration features such as:

• pull requests,
• code review,
• issues,
• access control,
• Actions automation,
• releases,
• web browsing of files and history.

Git can work without GitHub. GitHub uses Git, but it is not Git itself.

Other services can host Git repositories too.

Clone, pull and push

These three verbs are common:

• clone — make a local copy of a remote repository,
• pull — bring remote changes into your local repository,
• push — send your local commits to the remote repository.

If a tutorial says “git pull,” it is not downloading one random file. It is synchronizing Git history from a remote.

Why AI projects use them so much

Model code, configuration files, prompts, notebooks and deployment scripts all change over time.

Git lets a team record those changes. GitHub makes the repository easy to share and automate. The CLI lets you run the tools that actually train, test or deploy the project.

Together they form a basic working environment for modern AI development.

Do not paste commands blindly

Before running an unfamiliar command, check:

• what program it invokes,
• what files it will read or change,
• whether it requires administrator privileges,
• whether it downloads and executes remote code,
• which directory it runs in.

A terminal is powerful precisely because commands can make real changes.

One thing to remember

CLI is the command interface, Git is version control, and GitHub is an online platform for hosting and collaborating on Git repositories.

Lesson 028 introduces another word that appears whenever local AI starts using an NVIDIA GPU: CUDA.