Living Book Last updated June 2026

The First Hour with Claude Handbook

The whole mini-course, in one short read. Pop the kettle on, settle in, and go cover to cover — or dip into the chapter you need. It's kept up to date as Claude changes, and you can take a PDF copy with you.

How to read this book

This little handbook covers everything in the Your First Hour with Claude session — six short chapters, one per lesson. Read it front to back in well under an hour, or jump to the bit you need. Each chapter ends with a small thing to try, because the learning happens the moment you actually open Claude and have a go.

Claude changes quickly, so this is a living book: the online version is the source of truth, and the PDF is a snapshot. If your copy is more than a few months old, pop back for a fresh one.

Section I

Getting Started

Chapter One

Meet Claude

Let's begin gently. Claude is an AI assistant made by a company called Anthropic, and if you've used a chatbot before, the front door will feel familiar: you type something, it types back. But the desktop app has a little more to it than a single chat box, and knowing the shape of it up front makes everything that follows click into place.

Claude has three modes, and they sit as three tabs at the top of the app:

  • Chat — ask and answer. The familiar chatbot, and where you'll spend most of your time at first.
  • Cowork — Claude with hands. It can open and edit your files, use connected apps like your calendar and email, and even run jobs on a schedule.
  • Code — Claude as builder. For making and editing real software projects. The advanced end of town; we'll only peek at it.

There's one more thing worth knowing: Claude comes in a few sizes, called models. Opus is the most capable, Haiku is the fastest, and Sonnet sits in between. You can pick one from a little selector; most of the time the default is perfectly fine, and you can reach for Opus when you want maximum brainpower.

Key insight

The three modes form a simple ladder: the further right you go — Chat, then Cowork, then Code — the more Claude can do for you, and the more setup and trust it needs from you. You only need Chat to begin.

Put the kettle on and try this

Open the Claude desktop app (or claude.ai) and just find the three tabs — Chat, Cowork, Code. Click between them and watch the sidebar change. That's it for now; you've already got your bearings.

Chapter Two

Your First Chat

Chat is the front door, and the good news is there's no special language to learn — you type in plain English, exactly as you'd brief a sharp colleague. The quality of what comes back, though, depends almost entirely on the quality of what you put in. A vague ask gets a vague answer; a clear one gets something you can actually use.

Four habits that lift every answer

  • Give context — who it's for, what you're trying to achieve, anything relevant only you know.
  • State a specific goal — "draft a friendly two-line reply declining this" beats "help with my email".
  • Ask for the format you want — a list, a table, three options, a short paragraph. Say so.
  • Treat it as a conversation — follow-ups are free. "Shorter." "More formal." "Give me three versions."

Watch what a little context does. Before:

Beforewrite me a message about the school holidays

And after:

AfterHelp me write a short, warm message to other parents suggesting a few free school-holiday activities in the Blue Mountains for primary-aged kids. Friendly tone, three or four ideas, a sentence each.

One lovely thing about Claude: it's happy to take a position. Ask "which of these should I pick, and why?" and you'll get a clear recommendation with reasons — a verdict and a next step — rather than a wishy-washy wall of text. That's exactly what you want from a good assistant.

Key insight

Context plus a specific goal plus the format you want equals a great answer. You rarely nail it first time, and that's fine — refining is half the craft.

Put the kettle on and try this

Ask Claude something real from your week. Then send one follow-up that adds context or asks for a different format, and watch the second answer sharpen. That little loop — ask, refine — is most of what good prompting is.

Section II

Putting Claude to Work

Chapter Three

Claude Makes Things: Artifacts

Here's the first proper "oh!" moment. Claude doesn't only write text back at you — it can build a real, working little app or page, right there in the conversation. These are called artifacts, and they're where most people realise Claude is a maker, not just a talker.

In the session we build a "School Holiday Activities" planner: a tidy, filterable board of family outings, pulled together from messy real-world bits and pieces — dates, prices, locations, age ranges. And it genuinely works. The filters filter. The views switch. You can flip to a day-by-day calendar.

Why artifacts are special

  • It's a working app, not a screenshot — the buttons and filters actually do things.
  • It pulls scattered reality together into one useful place.
  • It's changeable on request — just ask.
  • It persists, so something you build once keeps earning its keep.

The magic trick is that you can keep talking to it:

Tweak it liveAdd a "free only" toggle, and show the Blue Mountains activities first.

Key insight

An artifact is real, working software made from a plain-English description. If you can describe what you want, Claude can often just build it.

Put the kettle on and try this

Ask Claude to build you a tiny tool for something you do — a packing checklist, a simple budget splitter, a weekly meal planner. Then ask it to change one thing. Watch it update in front of you.

Chapter Four

Claude on Your Computer: Cowork

So far Claude has been a clever voice in a box. In Cowork mode it grows hands. It can open and edit your actual files, use apps you've connected — your calendar, your email and more — and produce real deliverables on your own computer. You give it a task; it goes off, does the work, and reports back.

A lovely everyday example is a "morning brief": each day, Claude looks at your calendar and your inbox and hands you a tidy summary of what's on and what needs a reply. You wake up, and it's already done.

The shape of working with Cowork

  • It reads and writes files — real documents, in folders you point it at.
  • It uses connected apps — with your permission, and only the ones you allow.
  • You give a task and get a result back — a draft, a summary, a finished file.
  • You stay in charge: Claude drafts and does, you review and approve.
A first Cowork taskEach weekday morning, give me a short brief: today's calendar events, any urgent unread emails, and a three-item to-do list. Keep it to half a page.

Key insight

Cowork is Claude with hands. The same idea powers bigger projects too — it's exactly how you'd let Claude work through something like your taxes, file by file.

Put the kettle on and try this

In Cowork, connect your calendar and ask Claude for a one-paragraph summary of your week ahead. Notice how it does the work and hands it back — and that you get the final say on anything it produces.

Chapter Five

Claude on Autopilot: Scheduled Tasks

The second "oh!" moment: Claude doesn't only act when you ask. It can run jobs on a schedule — quietly, in the background, while you get on with your life. Set something up once, and Claude does it every morning, or every Monday, and leaves the result waiting for you.

The anatomy of a scheduled task

Every scheduled task is just two simple things plus a bit of bookkeeping:

  • A saved instruction — what you want done each time, written once in plain words.
  • A cadence — when it runs: daily, weekdays, or weekly.
  • It then runs automatically and leaves you the result.
  • And it keeps a history, so you can look back at every run.

Good first automations are humble and genuinely useful: a morning brief, a daily news digest on a topic you care about, or a weekly check of the supermarket specials so you never miss a deal on the things you actually buy.

A weekly scheduled taskEvery Monday morning, check the specials at my usual supermarkets for the items on my regular shopping list, and send me a short summary of what's on sale this week.

Key insight

A scheduled task is a saved instruction plus a cadence. It's the "set it and forget it" of AI — small, repeatable jobs that quietly do themselves.

Put the kettle on and try this

Set up one simple daily task — "each morning, summarise my unread email" is a perfect first one. Leave it overnight, and enjoy waking up to something already done.

Section III

Carrying On

Chapter Six

Where to Next

Look how far you've come in an hour. You've met Claude and its three modes, had a proper conversation, watched it build a working app, set it to work on your computer, and sent it off to run a job on its own. That's a genuinely capable toolkit — and the goal of this first hour was never to make you an expert, just to make you comfortable. Job done.

The one idea to keep

If you remember nothing else, remember the ladder: Chat answers, Cowork acts, Code builds — and the more capability you hand Claude, the more trust and setup it needs. Start with Chat, grow into Cowork as it earns your confidence, and peek at Code when you're curious.

Build the habit

The app's dashboard quietly tracks your activity — sessions, streaks, recent work — and it's a surprisingly good nudge for turning "I tried it once" into "I use it most days". Capability compounds for people who keep showing up; ten relaxed minutes a day will take you a long way.

Keep going with the series

This was the first of four free one-hour sessions. If you enjoyed it, the rest of the week builds on it:

  • Midjourney & video creation — making images and short video with AI.
  • Do your taxes with Claude Cowork — a real, end-to-end Cowork project.
  • Build something with AI in an hour — from idea to a working thing, a taste of Code mode.

You can sign up for the lot on the live sessions page, and everything from this session lives on in the interactive dashboard.

Key insight

Confidence beats expertise on day one. You now know enough to be genuinely useful with Claude — the rest is just happy practice.

Put the kettle on and try this — one last time

Pick one thing from this book and actually use it tomorrow: a sharper chat, a quick artifact, or a single scheduled task. One real habit is worth more than a whole hour of watching. Off you pop — and enjoy.

You've reached the end — of the book, not the hour

For the interactive version of everything here — the six lessons with screenshots and walk-throughs — head to the Your First Hour with Claude dashboard. And remember: this is a living book. Check back for the latest edition, or grab a fresh PDF whenever Claude has moved on.