Claude: a tour of the screen

Getting there and signing in

Claude lives on a website. Open your web browser and go to claude.ai. There's also a desktop app and phone apps, but the website is the simplest place to start and everything in this phase works there.

You'll be asked to sign in or sign up. You can create a free account with an email address, or sign in with a Google account if you already have one. Making an account is free, and you don't need to add a card to start. The sign-in screen is plain: a box for your email, and a button to continue with Google. Pick whichever you prefer and follow the prompts.

Once you're in, you land on a mostly empty screen with a box in the middle. That's home. There's far less going on here than it first appears, which is deliberate. The whole design pushes you towards one thing: typing a message.

The message box, and the basic loop

The box in the middle is where everything happens. You type your message (your prompt, the instruction or question you type in) and press Enter, or click the send button. Claude reads it and replies just below. You read the reply, type a follow-up, and round you go.

That's the entire basic loop: type, send, read, type again. Everything else on the screen exists to support that one action. If you remember nothing else from this tour, remember that Claude is a conversation, not a form. You're allowed to be chatty, to correct it, and to ask again.

Here is that loop with a real Fernway task, so you can see the shape of a first message rather than an abstract "type something":

Here is our company hybrid-working policy. In plain English, how many days a week are people expected to be in the office, and what are the core hours everyone should be contactable? Quote the exact lines you're relying on.

[paste the policy here]

Claude reads the pasted text and replies with the two facts and the lines they came from (the minimum office days and the core-hours window) so you can check them against the document in seconds. That "quote the exact line" habit is worth building from your very first message.

Starting a new chat

Look for a New chat button, usually near the top-left. Clicking it opens a fresh, blank conversation.

This matters more than it sounds. Claude remembers what you've said within a single chat, so a conversation about a holiday and one about a work report are best kept apart. Everything you type stays visible to Claude for the rest of that chat, which is helpful when you're building on one task, and muddling when two unrelated tasks share a window. Start a new chat when you move to a new topic, and things stay tidy.

Where your chats are listed

Down the left-hand side is a list of your past conversations. Each one is saved automatically and given a short title based on what it was about. Click any item to reopen it exactly where you left off. Nothing is lost when you close the tab. You can rename or delete these too, which is useful for clearing out the little practice chats you'll create while finding your feet.

So the main screen is simple: your past chats run down the left, and the message box sits in the middle. If the list is hidden (common on phones and narrow windows) there'll be a small menu icon to open it.

Attaching a file

This is one of the things Claude is best at, so it's worth finding early. Near the message box there's a way to attach a file, usually a paperclip or a plus icon. Click it, choose a document from your computer, and Claude can read it. It handles the everyday formats: a PDF, a Word document, a spreadsheet, a text file, or an image of a page.

You can also paste a large chunk of text straight into the box. Claude is comfortable with a lot of text at once. The amount it can hold in view during one conversation is its context window, so both attaching and pasting work well.

I've attached our weekly operations meeting notes. Give me a five-bullet summary a manager could read at a glance, then a separate list of every action someone agreed to do.

The plans, in plain terms

Claude has a free version and several paid plans. The exact names, prices and limits change from time to time, so treat the shape of it, not the figures, as the thing to remember, and check the official plans and pricing page for today's detail.

• Free: full chatting and a capable version of Claude, with limits on how much you can use in a given window.
• Pro: a personal paid plan that raises those limits and gives you more features and newer models sooner.
• Max: a higher personal tier with much more usage, for people who lean on Claude heavily.
• Team and Enterprise: plans an organisation buys for a group, adding shared administration, security controls and single sign-on.

Our honest advice: start free. It's a real, working tool, not a crippled demo, and you won't know whether you need more until you've used it a while. Only pay when you hit a specific limit that's actually in your way. And remember: a paid plan does not make answers immune to being wrong, and confidential work data belongs on an employer-approved tool, whatever plan you're on.

The model choice

Near the message box there's a small menu, the model picker, that lets you choose which version of Claude, or model, answers you. In plain terms the options sit on a single spectrum:

• A fast, everyday model: quick and light, ideal for summaries, simple questions and drafting. Most of the time this is all you need.
• The most capable model: slower and more thorough, built for hard reasoning, tricky analysis and long multi-step problems.

The individual model names (families such as Haiku, Sonnet and Opus) change as new versions arrive, which is exactly why it's worth thinking in terms of fast versus most capable rather than memorising a name. You may also see a newer frontier model listed with its own usage rules, where heavy use draws on separate usage credits beyond your plan's normal allowance; even then the same fast-versus-reasoning question applies, so pick by the job rather than by reaching for the newest name. Anthropic's guide to choosing a model keeps the current line-up.

Two reassuring things. First, if you're unsure, the default balanced option handles the large majority of everyday work; you don't have to get this right. Second, you can change the model partway through a conversation: pick a different one from the menu and your next reply uses it, while the chat keeps all the context above. So you might draft quickly with the fast model, then switch to the most capable one when you reach the part that needs careful thought, all in the same window.

That summary's great. Now, switching to the most capable model, read the whole policy again and tell me: are there any places where two clauses could contradict each other for someone who works fully remotely? Explain your reasoning step by step.

Try it now

Common mistakes

• Piling every topic into one endless chat. It gets muddled, because Claude can see everything above. Start a new chat for each new topic.
• Typing out a long document by hand. You don't need to; attach the file or paste the text.
• Assuming you must pay to try it properly. You don't. The free plan is capable; only pay when a specific limit blocks you.
• Chasing the "best" model for everything. The most capable model is slower and burns through your usage faster. For summaries and drafts the fast one is fine, so save the heavy model for hard reasoning.
• Over-trusting a confident first reply because the tool looks slick. A tidy interface and a fluent answer are not the same as a correct one. Ask Claude to quote its source lines and check anything that matters. The polish is in the writing, not a guarantee of the facts.

Keeping current

Claude's screen, plans and models change more often than almost anything else in this course: new models arrive, limits shift, features move. The durable skills (type, attach, check the quotes) stay put; the specifics won't. One specific worth knowing: the Claude 5 generation of models arrived in June 2026, and access to the newest frontier model tends to move onto usage credits once you go beyond a plan's included allowance, so the picker and its usage rules may look different from this tour. When something looks different, check the official Claude release notes and the help centre. Accurate as of 14 July 2026.