Getting Started with Claude: The Non-Developer's Setup Guide

Claude has changed a lot in the last few months: memory on every plan, a new agentic desktop mode, free voice, and a steady stream of interface updates. If you've been meaning to set it up properly, this is the no-jargon guide. Three ways to access it, one account, and everything you need to get from "signed up" to "getting value."

1. Three ways in, one account. Web (claude.ai), a native desktop app (Mac/Windows), and mobile apps (iOS/Android). Sign in once; your chats, projects, and memory sync everywhere.

2. Pricing today: Free $0 · Pro $17/mo billed annually ($200 up front) or $20/mo monthly · Max from $100/mo. Plans and prices are subject to change.

3. What's new and worth knowing: Memory now works on every plan including Free; Claude Cowork (hands-off, agentic desktop work) is generally available; voice mode is free on web and mobile.

4. Files are generous: 500MB per file and up to 20 files per chat; project knowledge files are 30MB each with no count limit.

5. For marketers: the single best move is one Project per brand or client, loaded with your voice guide and best examples, producing reusable Artifacts.

The five most authoritative, currently-accurate setup guides. Official Anthropic documentation is the system of record (several were updated within the last week), plus one trusted independent starter for your very first steps.

1. Get Started with Claude — Anthropic Help Center
The canonical entry point covering web, desktop, and mobile: what Claude is, the three access paths, age (18+) and region requirements, how the model selector works, and free-plan usage limits. Updated this week. Start here.

2. Install Claude Desktop — Anthropic Help Center
The definitive desktop reference: system requirements, the exact install flow, Desktop Extensions (one-click local connectors), and Claude Cowork. Updated April 24, 2026.

3. Install Claude for iOS / Android — Anthropic Help Center. Where to download, supported OS versions, why the app might not appear (region/device), and the uninstall + subscription-cancellation caveats. Updated mid-May 2026.

4. Release Notes — Anthropic
Not a tutorial, but the authoritative, dated changelog. The fastest way to confirm whether any feature claim is actually current. Entries through early June 2026.

5. Claude: A guide to Anthropic's AI models and chatbot — Zapier
A friendly starting point if you're brand new. In plain language, no jargon: how to sign up at claude.ai, what you can do as a free user, and how the models differ.

Claude has one free tier and several paid ones. For an individual non-developer, the real decision is Free vs. Pro to Max and the team plans are for heavy or multi-person use.

• Free — $0. Best for trying Claude out. Includes: chat on web, iOS, Android & desktop; web search; memory; file creation & code execution; connectors; extended thinking.

• Pro — $17/mo billed annually ($200 up front), or $20/mo month-to-month. Best for everyday individual use. Adds: more usage, Claude Code, Claude Cowork, unlimited Projects, Research, more models, and the Microsoft 365 / Outlook add-ins.

Max — from $100/mo• (choose 5× or 20× Pro's usage). Best for power users who keep hitting Pro's limits. Adds: higher output limits, priority access at busy times, and early access to new features.

• Team — $20/seat/mo annual ($25 monthly); premium seats $100 ($125 monthly). Best for teams of 5–150. Adds: central billing, SSO, admin controls, and enterprise search.

• Enterprise — custom ($20/seat + usage at API rates). Best for large organizations. Adds: SCIM, audit logs, advanced security, and a HIPAA-ready option.

Prices exclude applicable tax. Plans and prices are subject to change at Anthropic's discretion. "More usage" is deliberately not a fixed number of messages, which is exactly why overages matter.

Claude doesn't bill you by the message. Each plan gives you a usage limit (think of it as a conversation budget) that refills on a schedule. What happens when you hit it depends entirely on your plan.

The basics of usage limits. Your limit is affected by how long and complex your chats are, which model you pick, and how hard you ask Claude to "think." A session limit resets roughly every 5 hours, and paid plans also carry weekly limits. One thing power users miss: your activity across claude.ai, the desktop app, and Claude Code all draws from the same limit.

On the Free plan, there are no overages. When you reach the limit, you simply wait for it to reset; nothing to buy, no surprise charge.

On paid plans (Pro, Max 5×, Max 20×), overages are opt-in and called "usage credits." By default, hitting your limit just pauses you until it resets. But you can switch on usage credits to keep working without interruption:

One more distinction: don't confuse "running out of usage" with "this chat got too long." Running out of usage is about your time-based budget: wait, upgrade, or use credits.

Pick the right model for the job. Which model you use is the single biggest lever on how fast you burn through your usage, so match it to the task:

You switch between them with the model selector right below the input box.

This is where most people should start.

1. Create your account. Go to claude.ai and sign up with an email or Google account. You'll need to be 18+ and in a supported region. That same login works on every device.

2. Pick a plan. The Free plan does real work, so start there — you can always upgrade once you hit its limits. See Pricing & Plans above for the full breakdown of Free vs. Pro vs. Max.

3. Learn the interface. Type into the box and hit send. The model you're talking to is shown right below the input box; click it to switch models or adjust how hard Claude "thinks." The "+" button (lower-left) or typing "/" opens more options. On the free plan, usage is session-based and resets every five hours.

4. Upload files. Click "+" → "Add files or photos," or just drag and drop. You can attach up to 500MB per file and 20 files per chat; PDFs, Word docs, spreadsheets, images, and more. PDFs under 100 pages are read for both text and visuals.

5. Set up Projects. In the left rail, open Projects → "New Project." Drop reference material (brand guide, past work, data) into the project knowledge base on the right, and I always "Set project instructions" to give Claude persistent rules. Everything in that project carries the same context. Free users get up to 5 projects; paid plans are unlimited.

The desktop app is worth installing on day two.

1. Download it. Go to claude.ai/download. It runs on macOS 11+ and Windows 10+ (there's no Linux build). Open the installer, launch the app, and sign in with your account. The free plan works fine in the app.

2. Know what it adds over the browser. The desktop app lives in your dock or taskbar with an instant-access shortcut from anywhere on your computer. Through Desktop Extensions (Settings → Extensions), it can connect to your local files, calendar, and apps with one click.

3. The agentic layer: Cowork or Claude Code. The desktop app is also your door to Claude doing hands-off, multi-step work: synthesizing research, organizing files, building a spreadsheet or deck from messy inputs. There are two doors to the same engine.

Cowork is the no-terminal one — click to install skills, but it only runs while your computer is awake and the app is open. I don’t recommend it for most folks.

Claude Code is the more powerful door, and the one we build on: it lets you create reusable skills, run scheduled jobs that fire even when your laptop is closed, and meet the governance of real client work demands. It asks a bit more of you to get started; but that's where the leverage compounds.

Which door fits you? We break it all the way down in:

Turn AI Into Your Most Powerful Marketing Teammate

Put AI to Work Across Your Marketing Strategy. In this hands-on course, you will learn how to confidently use tools like ChatGPT, Midjourney, and Descript to plan, create, and scale content across copy, images, and video.

You will leave with practical frameworks and real experience generating marketing assets you can apply immediately. Led by Alec Newcomb, this course is built for marketers, leaders, and business owners who want to improve efficiency, drive growth, and stay ahead with AI.

1. Install the app. Search "Claude by Anthropic" in the App Store or Google Play and install it (free). iOS needs version 17.0 or later. Sign in with your account to sync everything.

2. Use your voice. Open a chat and tap the sound-wave icon next to the microphone, pick a voice, and start talking. There's a hands-free mode (responds to natural pauses) and a push-to-talk mode (better in noisy places). You can switch between voice and text mid-conversation, and transcripts are saved.

3. Use the camera. Snap a photo or pull one from your gallery — a menu, a whiteboard, a competitor's ad, a receipt — and ask Claude about it.

4. Connect to your phone's apps. On mobile, Claude can help draft Messages and Mail, find places in Maps, and create Calendar events and Reminders (iOS).

5. Know the limits. Cowork itself runs only on the desktop app; your phone can assign a task to your desktop, but can't run Cowork on its own. Local file access and a few iOS integrations are narrower on mobile, and voice mode drains the battery faster.

Installing Claude is the easy part. What actually decides whether you get generic "AI slop" or something genuinely useful isn't a setting, it's how you talk to it. Most people type three words at Claude like it's a search box and conclude AI "just isn't that good." The people who get great results brief it like a teammate.

That's the whole idea behind our T.A.L.K. Method — four letters, no prompt-engineering course required:

• T — Talk, don't type. Use the voice mode you set up on mobile and talk the task through, the way you'd brief a colleague.

• A — Add your context. Tell Claude who you are, who it's for, and your rules ("COO at a B2B SaaS, writing for a recruiter skimming 200 applications, no buzzwords, keep it to one page").

• L — Lead with what "good" looks like. Paste an example to match, name the real outcome, and add "mark anything you're not sure about with [VERIFY]."

• K — Keep what works. Save the instructions, not just the answer — which is exactly what Projects are for. Set one up once and it loads every time.

Notice the overlap with everything above: voice mode is "T," and Projects are "A" and "K." Setup and skill are the same move.

The full walkthrough, with a five-minute starter and a downloadable PDF — lives here:

1. One Project per brand or client. Load it with your style guide, tone rules, banned words, brand colors, fonts, and three to five of your best examples, then write project instructions for how Claude should write. Every chat inherits it, no re-pasting your brief and it saves context too.

2. Treat Artifacts as reusable assets. Have Claude build drafts, comparison tables, content calendars, or simple interactive tools as Artifacts you can edit in place and embed; much better than copy-pasting out of the chat.

3. Repurpose in one pass. Claude can create and edit Word docs, spreadsheets, slide decks, and PDFs directly. Turn one long-form piece into a deck outline, a spreadsheet of social hooks, and a one-pager in a single thread.

4. Feed it real material. Drop in competitor PDFs, brand decks, or a CSV of survey data (500MB per file, 20 per chat) and ask for positioning analysis, message extraction, or audience insights.

5. Use it for competitive and creative review. Paste a competitor's ad screenshot or a landing-page grab and ask Claude to critique the hook, hierarchy, and call-to-action — or to reverse-engineer the messaging framework.

6. Turn on web research and connectors. Enable web search and Research for cited market scans, and wire up Google Drive or Slack connectors so Claude pulls from your real documents instead of guessing.

7. Let memory carry your context. Memory means Claude remembers your brand and ongoing campaigns across chats. Use Incognito chats for confidential client work you don't want retained. Projects to keep Memories in swimlanes.

8. Delegate the busywork to Cowork/Claud Code. Point it at a connected folder and have it batch-rename assets, assemble a formatted report, or compile research overnight while you do higher-value work.

9. Save your winning prompts. Keep a prompt-library doc inside each Project so your best brief patterns are one click away. But I use prompts less and less as the models get better, just talk to it.

10. Brief Claude like a teammate, not a search box. Spell out the audience, the task, the must/avoid constraints, and an example to match; that's the T.A.L.K. Method in miniature. It beats one-line asks every time.

Download the PDF version with setup checklists, plan comparisons, and the T.A.L.K. framework for future reference.

AI helps you speed up; that's the easy part. The part that actually matters: don't settle for someone else's setup.

The default Claude is a starting line, not a finish line. Keep tweaking yours; your voice, your examples, your rules, your Projects — until it stops feeling like a tool and starts feeling like a partner who knows how you work.

That's the whole game. Make it yours.

— Alec

Share with a coworker so they can join the MarketingAlec newsletter for practical, no-jargon AI guides built for marketers.

Subscribe free → https://ai.marketingalec.com