What is the Claude Code CLI? Plain English Explanation

by

You keep seeing “CLI” in Claude Code documentation and tutorials. Maybe you’ve already installed Claude Code and you’re wondering what the CLI actually is versus the desktop app or the browser version. Maybe “CLI” just sounds intimidating and you want someone to explain it plainly.

Either way, this article covers it.

This is part of our complete Claude Code Beginner’s Guide.

TL;DR

  • CLI** stands for Command Line Interface — it means a program you use by typing commands in a terminal
  • The Claude Code CLI is the claude command you run in your terminal
  • It’s the primary and most powerful way to use Claude Code
  • You don’t need to memorize many commands — most interaction is plain English conversation
  • A handful of slash commands (/help, /clear, /exit, etc.) handle session management

What Does CLI Mean?

CLI stands for Command Line Interface. It’s the technical name for any program you interact with by typing text commands in a terminal window, rather than clicking buttons in a graphical interface.

Think about the difference between these two ways to delete a file:

  • Graphical interface (GUI):** Find the file in Finder or File Explorer, right-click it, click “Move to Trash”
  • Command line interface (CLI):** Type rm filename.txt in the terminal and press Enter

Both delete the file. The CLI version is faster once you know the command. The graphical version is easier to discover for the first time. Most developer tools offer both options.

Claude Code’s primary interface is a CLI — you type commands and interact with it through text.

What is the Claude Code CLI Specifically?

When you install Claude Code, you get a program called claude on your computer. The Claude Code CLI is that claude program — the thing you launch from your terminal.

When you type claude in your terminal and press Enter, the Claude Code CLI opens an interactive session. From there, you type natural language instructions (not code or special commands) and Claude Code responds and acts.

The CLI is where Claude Code lives for most developers. It’s not the only way to use Claude Code — there’s also a Desktop app and IDE integrations — but the CLI is the foundation everything else is built on.

How the CLI Session Works

When you run claude in a project folder, a session starts. Think of it like opening a chat window — except this chat window has tools that can read and write files on your computer.

A session has three parts:

1. Your messages. You type instructions in plain English. “Add error handling to this function.” “Explain how the authentication flow works.” “Create a new API endpoint for user profiles.”

2. Claude Code’s actions. Claude Code reads files, writes changes, runs terminal commands, and reports what it’s doing. You see this happening in real time.

3. The conversation history. Within a session, Claude Code remembers everything you’ve said and done. You can refer back to earlier context without repeating yourself. This history doesn’t persist to a new session — when you close and reopen, you start fresh.

How to Launch the Claude Code CLI

Interactive mode (most common)

Navigate to your project folder and run:


# Start an interactive Claude Code session in the current folder
claude

This opens a session where you can have a back-and-forth conversation.

Single-command mode

For quick, one-off tasks, you can skip the interactive session entirely:


# Run a single task and get the output without entering interactive mode
claude "Add a docstring to every function in utils.py"

Claude Code runs the task, shows you what it did, and exits. Useful for scripting or quick jobs.

Start with a specific file in focus

You can open a session with Claude Code already having read a specific file:


# Start a session and load a specific file into context immediately
claude --file src/app.js

The Commands You’ll Actually Use

Most of your time in Claude Code is spent typing natural language — not slash commands. But there are a handful of built-in commands worth knowing.

/help


# Show a list of available commands and keyboard shortcuts
/help

Prints a help menu. Use this whenever you’re not sure what’s available.

/clear


# Reset the conversation without exiting Claude Code
/clear

Clears the conversation history and starts fresh within the same session. Useful when you switch to a completely different task and don’t want the old context influencing new responses.

/exit


# Exit the Claude Code session
/exit

Closes the session cleanly. You can also press Ctrl + C twice.

/config


# Open Claude Code's configuration settings
/config

Opens an interactive configuration menu where you can adjust settings like the auto-update channel, model preferences, and other options — without editing a settings file manually.

claude --version

Run this outside of a session to check which version of Claude Code you have installed:


# Print the installed version of Claude Code
claude --version

claude doctor

Run this outside of a session to check the health of your installation:


# Run Claude Code's built-in installation diagnostics
claude doctor

Prints a status report covering your installation, authentication state, settings, and any detected issues.

claude update


# Manually check for and apply updates to Claude Code
claude update

The native installer auto-updates in the background, but if you want to trigger an update immediately, this command does it. For more on updates, see How to Update Claude Code to the Latest Version.

CLI vs Desktop App vs IDE Extension

The Claude Code CLI is the terminal-based interface. But Claude Code also runs in other environments:

Desktop App: A graphical application that gives you a Claude Code interface without typing terminal commands. Good for users who prefer not to use the terminal. Available for Mac, Windows, and Linux.

VS Code Extension: Claude Code integrates directly into VS Code, so you can use it alongside your editor without switching to a terminal window.

JetBrains Plugin: Same idea — Claude Code inside JetBrains IDEs like IntelliJ, PyCharm, and WebStorm.

All of these use the same underlying Claude Code. The CLI is the most powerful and flexible version because it gives you full access to every feature. The other interfaces are built on top of it.

💡 Pro Tip: Many developers run the CLI in a terminal panel inside VS Code (View → Terminal) and use the VS Code extension at the same time. You get the full CLI and the editor integration together.

Keyboard Shortcuts in the CLI

The Claude Code CLI session has a few keyboard shortcuts that make navigation faster:

Shortcut What it does
**Ctrl + C** Stop the current operation or exit the session
**Up arrow** Scroll through previous messages
**Ctrl + L** Clear the terminal display (doesn’t reset conversation)
**Enter** Send your message
**Shift + Enter** Add a new line without sending (for multi-line messages)

For the full shortcuts reference, see Claude Code Keyboard Shortcuts — The Complete List.

Why Use the CLI Instead of the Desktop App?

Both work. For most power users, the CLI is preferred because:

  • It’s faster to launch (cd project && claude takes two seconds)
  • It integrates naturally with other terminal tools and scripts
  • It works on remote servers via SSH
  • It’s what most Claude Code documentation and tutorials assume

If you’re new to terminals and find the graphical Desktop app more approachable, start there. You can always switch to the CLI later — it’s the same tool underneath.

Common Mistakes with the Claude Code CLI

Running claude from the wrong folder

Claude Code reads the folder you’re in when you start the session. If you run claude from your home directory instead of your project folder, it won’t find your project files and won’t understand your codebase. Always cd into your project first.

Confusing claude (the CLI) with claude.ai (the website)

They use the same AI but are different products. When this guide says “run claude,” it means type claude in your terminal. It doesn’t mean open a browser.

Expecting the CLI to remember previous sessions

Claude Code doesn’t have persistent memory across sessions by default. When you close a session and start a new one, the conversation starts fresh. Give it a quick overview at the start of each new session if the context matters.

Using sudo with claude commands

Never run claude with sudo. Claude Code doesn’t need elevated permissions to work, and using sudo can cause file permission issues that are annoying to fix.

Key Takeaways

  • The Claude Code CLI is the claude command you run in your terminal
  • CLI** just means “text-based interface” — there’s nothing more mysterious about it
  • Most interaction is natural language, not code commands
  • A few slash commands (/help, /clear, /exit, /config) handle session management
  • claude doctor and claude --version are the most useful diagnostic commands to know
  • The CLI is the most powerful way to use Claude Code, though Desktop and IDE options also exist

Next Steps

Return to the Claude Code Beginner’s Guide for the full learning path.