Where Claude Code Lives — and Where You Should Start
Here's the moment most people quietly give up. They hear Claude Code, they go to try it, and the very first thing a tutorial puts in front of them is a black screen with a blinking line — waiting for them to type a command they've never seen. And the thought lands almost instantly: this is a developer thing, not a me thing.
So before you picture that black screen one more time, take a breath. The part those tutorials skipped is the part that changes everything: you don't have to start there at all.
You don't need the terminal
That black screen has a name — the terminal — and yes, it's one way in. It is not the only way, and it's nowhere near the friendliest. It's the door tutorials love to show first, which is exactly why so many beginners bounce off in the first minute.
Claude Code also comes as a normal app you open like any other, and as a web page you sign into in your browser. Same Claude. No commands to memorize, nothing to install from a command line. You click, you sign in, you're looking at a screen — the way you'd expect any app to work.
Same Claude, several doors
Picture Claude Code as one room with several doors into it. There's the terminal. There's your code editor, if you happen to have one (the popular ones are VS Code, Cursor, and JetBrains). There's a desktop app. And there's the web, right in your browser.
Walk through any of those doors and you end up in the same room, with the same Claude, doing the same things. The door only changes the chair you sit in — not who's in the room. So you get to pick the one that feels easiest, and for almost everyone starting out, that's the desktop app or the web.
Already at home in a terminal or in VS Code? You're not stuck with the beginner's door — we open the others properly later in the course. For now, the app keeps things simple.
Let's get you in
The lightest path takes about a minute, and there's nothing to memorize.
Open the Claude desktop app, or go to the Claude Code page in your browser. Either one is fine — same room, remember.
Sign in with your Claude account — the same login you'd use for the regular Claude chat. Nothing new to set up.
Switch to the Code part of the app. You'll know you're in the right place when you see a big, friendly box inviting you to describe a task or ask a question. That box is the heart of the whole thing.
(The exact button labels shift around as the app gets updated — so go by what things do, not the precise wording. If a name here looks slightly different on your screen, you're still in the right place.)
A friendly tour of the screen
At first glance, the screen has a lot on it — little labels and icons tucked into the corners, a panel up top, buttons hugging that text box. It can look like a cockpit. Here's the secret nobody says out loud: on day one, you'll touch maybe four of these. Everything else is just there, waiting politely for the lesson where it matters.
Let's name all of it anyway, so nothing on the screen is a stranger.
Going tag by tag, here's every marker on the screen. The only one you truly need today is number 3 — the box where you type. The rest you're just meeting in passing.
- New session — a clean slate. Click it whenever you want to start over from scratch.
- Activity panel — a quick look at what you've been up to lately, including a playful pixel-grid of your recent sessions.
- The prompt box — describe a task or ask a question. This is the one you live in: type what you want, in plain English, and press enter. Learn only this control and you could already use Claude Code.
- Auto (the mode control) — sets how much freedom Claude has to act before it checks back with you. The one thing to know now: out of the box, Claude asks before changing anything. It gets a whole lesson of its own later.
- Microphone — speak instead of type, if you'd rather talk it through.
- The "+" button — add things to the conversation: a file, a photo, a screenshot, and Connectors, which let Claude reach other tools. More on those later in the course.
- Local (the environment) — where the work runs. "Local" just means right here on your own computer; it can also run elsewhere.
- The project folder — the folder Claude is working inside (here it reads "agent-hub"). Pointing it at a folder is exactly what we do next.
- main (the branch) — which version-control "save-line" you're on. Fine to leave alone for now.
- worktree — an advanced Git feature. As a beginner, you can completely ignore it.
- The folder button — points Claude at a different folder. We'll use this next lesson too.
- The model — which Claude brain you're using. There are faster ones and deeper ones; the default is a fine place to start.
- Effort — how hard Claude thinks, on a simple faster-versus-smarter setting. Also fine left as-is.
- The pixel creature — just Claude Code having a bit of personality. It doesn't do anything. It's there to make you smile.
Nothing here is a mystery now
That's the whole cockpit. Not one mystery control left — a box you type in, a few settings hugging it, a strip that says where Claude's working, and a couple of dials in the corner you can grow into. The black-screen terminal you were bracing for? You never had to touch it.
You can name everything you're looking at now. Next, we put it to work: we'll point Claude Code at a real folder on your computer and ask it for its first actual change — the moment it stops being a screen and starts being useful.