Your First Five Minutes: Installing and Running Claude Code
This lesson assumes you have never opened a terminal in your life. That's not a caveat. It's the actual starting point. Everything from here to the end of this course happens inside the tool this lesson installs, so nothing after this point works without it.
What a terminal actually is
Instead of clicking icons, you type a line, press Enter, and read the reply as text. That's the entire idea. It looks intimidating because it's unfamiliar, not because it's hard.
Every Mac and Windows computer already has one built in, so no download is required to get the terminal itself.
Open Spotlight (press Cmd + Space), type Terminal, press Enter. A window opens with a blinking cursor. That's it: that's the terminal.
A blank terminal window is not broken. It has no icons, no menus, and no hint of what to do, because everything happens by typing a line and reading what comes back. That absence of visual furniture is normal, not an error state.
Installing Claude Code
Claude Code installs with one line typed into that terminal window.
curl -fsSL https://claude.ai/install.sh | bash
Use the block above, matching macOS/Linux or Windows. Right-click (or Cmd+V / Ctrl+V) to paste into the terminal window, since typing a URL like this by hand invites typos.
Text scrolls past. This is the installer downloading and unpacking Claude Code onto your machine. It typically takes under a minute on an ordinary connection.
Type claude --version and press Enter. Seeing a version number back (rather than an error like "command not found") means the install succeeded.
On macOS, this usually means the terminal needs restarting. Close the window and reopen it, then try claude --version again. On Windows, it can mean PowerShell needs to be reopened as well. This is a normal first-run hiccup, not a sign the install failed.
Your first run
Type claude and press Enter. A prompt appears, waiting for a line of plain English: no special syntax, no punctuation rules, no programming.
Type this exactly, as a test:
What files are in this folder?
What happens next is the entire idea of this course in miniature: Claude Code reads the actual folder on your actual computer and answers with what it actually found there: not a guess, not a generic answer, a real look at real files. If the folder is empty, it says so. That's the agent behaving correctly, not the tool failing.
What "typing plain English into it" looks like from here
Every worked demo in this course (the spreadsheet reconciliation, the contract folder, the monthly report) starts exactly the way this first run did: open a terminal, type claude in the folder that holds the relevant files, then type a sentence describing the job. There is no additional interface to learn. The screen stays text, the input stays a sentence, and the only thing that changes lesson to lesson is what that sentence asks for and how carefully it's written. That's what practice 01, two lessons from now, is actually about.
Open the terminal inside the specific folder you want Claude Code to work on, not your whole computer's home folder. On macOS and Windows, most file-browser apps have a right-click option like "Open in Terminal" on a folder, which starts you in exactly the right place. Everything ahead in this course assumes Claude Code is pointed at a specific, bounded folder, not "everywhere."
What comes next
The terminal and the install are done. The same window and the same "type a sentence, read the reply" pattern carries through every lesson from here on. Next: the two doors into this same agent, and why the professional one is where real business data belongs.
Claude Cowork vs. Claude Code: the same engine, two very different doors, and which one this course is built on.
Reply here and it goes straight to Rod. Same as replying to one of his emails.
