What it is: The 2026 practical guide to using Claude AI: how to access it (web, desktop, mobile, API), which model to use when (Sonnet 5.5, Opus 5.5, Haiku 5.5 or Fable 5.1), what Claude does best and where it falls short, free vs. paid, and the prompting habits that produce better answers. Sourced from Anthropic’s official documentation.
Who it is for: Anyone trying Claude for the first time or trying to get more value from a Claude Pro subscription they already pay for.
Best if: You want a clear, current snapshot of what Claude can do in 2026 and how to actually use it day-to-day.
Skip if: You’re a developer. Jump to our Claude Code guide. For one practical AI workflow each morning, subscribe to our free daily newsletter.
Claude is an AI assistant built by Anthropic, and if you’ve only ever used ChatGPT, the first thing you’ll notice is that Claude feels different. It tends to write more naturally, reason more carefully, and give direct answers instead of hedging. Whether you’re new to AI entirely or comparing tools, this guide walks you through exactly how to use Claude: what it can do, how to access it, and how to get the best results from day one.
Short version: go to claude.ai, create a free account, and start typing. That’s it. But the real value comes from understanding what makes Claude different and how to prompt it well, which is what the rest of this guide covers.
Updated October 2026: This guide now covers Claude’s current models (Sonnet 5.5, Opus 5.5, Haiku 5.5 and Fable 5.1), the new Claude Duet app, memory across chats, web search on every plan, and the privacy setting that decides whether your chats are used to train Claude. See our Claude Opus 5.5 release notes and Opus vs Sonnet vs Haiku comparison.
What is Claude actually?
Claude is a family of large language models made by Anthropic, a San Francisco-based AI safety company founded by former OpenAI researchers. Like ChatGPT and Gemini, Claude is a conversational AI: you type a question or request, and Claude writes back. But Claude has its own personality. It tends to give clearer answers and hedge less when it knows the answer.
Claude is used for writing, editing, summarizing long documents, analyzing data, brainstorming, research, coding, customer support, and pretty much any task that benefits from structured thinking or clear writing. According to Anthropic’s documentation, Claude excels at “tasks involving language, reasoning, analysis, coding, and more,” which in practice means it’s one of the most versatile AI tools available in 2026.
How do you access Claude?
There are four main ways to use Claude, depending on what you’re trying to do:
1. claude.ai (the web chat)
Most people start at claude.ai. Sign up with your email or a Google account, and you’re talking to Claude in under a minute. It works like any chat app: type a message, get a reply, and keep the conversation going. The free plan is enough for light use.
2. The desktop and phone apps
Claude also has apps for Mac and Windows and for iPhone and Android. They all use the same account, so your chats follow you from one to the other. Voice mode, where you talk and Claude talks back, works on every plan, including Free. The desktop app can also work with files on your computer. See our Claude desktop app guide.
3. Claude Code (for developers and builders)
If you want Claude to build software, editing files and running commands on your computer, that’s Claude Code, Anthropic’s coding AI agent. It reads your project, edits files, runs commands and checks its own work. You can use it in your computer’s terminal, a code editor, the desktop app or the browser, and it needs a paid plan.
4. The Claude API (for developers building apps)
If you’re building a product or integration, Anthropic offers a Claude API and a Claude Agent SDK available in Python and TypeScript. This guide doesn’t cover the API in depth. For that, see our Claude Agent SDK tutorial.
What are the Claude models (Opus, Sonnet, Haiku)?
Claude comes in several models. As of October 2026 there are four current ones, each with different strengths. On claude.ai, you choose which one to use from the model menu in the chat interface.
- Claude Sonnet 5.5 (September 2026): the best mix of speed and intelligence, and the one most people use day to day. Available on every plan, including Free. Details from Anthropic.
- Claude Opus 5.5 (September 2026): built for long, demanding work like complex analysis, big documents and coding projects it runs step by step. Available on Pro and higher plans, not on Free. Learn more from Anthropic.
- Claude Haiku 5.5 (October 2026): the fastest and cheapest model, for quick answers and high-volume tasks. Available on every plan.
- Claude Fable 5.1 (September 2026): Anthropic’s most capable model, part of its Mythos family, for the hardest reasoning and longest jobs. Paid plans only: Max includes it within your weekly limits, and Pro uses pay-as-you-go usage credits. How Fable works on each plan.
If you’re unsure which to use, start with Sonnet 5.5. If the task is complex or the quality matters a lot, switch to Opus 5.5. If you need speed above all else, use Haiku 5.5. Save Fable for the hardest problems, since it uses your limits fastest. You can change models mid-conversation without losing context. Anthropic maintains a full models overview with technical details.
What should your first 5 minutes with Claude look like?
- Go to claude.ai and sign up for a free account.
- Pick a real task. Don’t just type “hi.” Give Claude something you actually need help with: a paragraph to rewrite, a document to summarize, a question to brainstorm around.
- Tell Claude who it should be. “Act as an experienced editor.” “Be a patient teacher.” This single tweak dramatically changes the output.
- Include context. Paste the source material, describe your audience, explain what you’ve already tried. The more Claude knows, the better the response.
- Iterate. The first response is rarely perfect. Tell Claude what to change (“make it shorter,” “more casual,” “cut the intro”) and it will.
Most people hit Claude once, don’t love the response, and walk away. The people who actually get value from Claude treat it like a conversation, not a search box.
What does using Claude look like day to day?
Here are three everyday jobs you can hand Claude on the Free plan, drawn the way they look in Claude: your request at the top, then the steps Claude takes. Copy any of the requests and change the details to fit your week.
Turn your bank statement into a budget
Download last month’s transactions from your bank as a spreadsheet file and attach it. Claude sorts every line into a category, adds up the totals and hands back a spreadsheet you can open in Excel or Google Sheets.
Works on the Free plan| Category | Spent | vs. August |
| Groceries | $612 | +$48 |
| Eating out | $284 | +$91 |
| Subscriptions | $137 | same |
| Gas | $166 | −$22 |
Find the emails that need you
Add the Gmail connector from the Connectors menu in Claude (our Gmail setup guide shows each click), then ask Claude to go through your inbox. It gives you a short list of the ones waiting on a reply, with a draft for each, and it asks for your approval before anything is sent.
Works on the Free planPlan your week around your calendar
Add the Google Calendar connector the same way. Claude can see when you’re free and fit the things on your list into the gaps, so the jobs you keep putting off finally get a slot.
Works on the Free planIllustrations drawn by Beginners in AI to show what a session looks like. Your own steps and numbers will differ.
What is Claude especially good at?
Writing
In our use, Claude writes the most natural-sounding prose of the big AI assistants. It’s less formulaic and less prone to the usual AI tics (overusing “furthermore” or “it’s important to note”). Paste in three samples of your own writing and ask Claude to match the voice, and it will pick up your style closely.
Long Documents
Claude’s current models can work with up to 1 million tokens of context, roughly 750,000 words (in the app, the exact limit varies by model). You can drop in a 500-page contract, a year of meeting notes, or a book and ask Claude to analyze it. Gemini also handles very long documents well, so it’s worth trying both on your own files.
Coding
Claude is a favorite for serious coding work because it reasons through problems, makes careful edits, and has its own Claude Code tool for building software.
Thoughtful Analysis
When you ask a hard question with no clear answer, Claude works through the trade-offs instead of giving you a safe middle-ground answer. That helps most with big decisions, business strategy and ethical questions.
What can Claude not do well?
- No photo or illustration generation. Claude can analyze images you upload and build charts, diagrams, designs, slide decks and documents (Claude Design, Slides and Docs are free to use). For photos and illustrations, use ChatGPT or a dedicated tool like Midjourney.
- Fresh information needs a search. Claude has a knowledge cutoff, but web search works on all plans, Free included, so it can look up current news and prices when you ask. Research mode, which builds longer reports from many sources, needs Pro or above.
- Fewer ready-made add-ons. ChatGPT has a bigger catalog of apps and plugins, and Gemini is built into Google’s apps. Claude has caught up fast with MCP (a standard way to plug other apps into Claude), Claude Skills, connectors and a plugin directory, but there’s still less to choose from.
What is new in Claude in 2026?
If you last used Claude a year ago, a lot has been added. Here’s what’s new.
Claude Duet (new in October 2026)
Claude Duet is a separate app, launched October 21, 2026, where Claude works like a personal assistant that checks in on its own. You message it like a person, by text or voice. It sends you a morning brief, checks your connected apps at set times, and takes on jobs like setting up meetings, drafting briefs and decks, or building a dashboard. It runs on its own computer in the cloud, so it keeps working after you close the app, and it asks you before it sends a message, makes a payment, signs a document or deletes anything.
Who can get it: Pro and Max users in the US, through a waitlist, as a research preview. It costs nothing extra and uses your plan’s usage limits. It runs on Opus 5.5, and you can’t switch models.
How to start: join the waitlist on the Claude Duet page using the email on your Claude account, then pick your plan and phone type (iPhone or Android). When your invite email arrives, install the app (or use Duet on the web at claude.ai) and sign in with your Claude account. Anthropic’s getting-started guide lists what it can’t do yet.
Memory
Claude remembers useful details across your chats, so you don’t have to re-explain your job or your preferences every time. Every plan has memory, and on personal plans (Free, Pro and Max) it’s switched on by default. To see what it knows, go to Settings > Memory, where you can edit or delete any item. Health, beliefs and other sensitive topics stay out of memory unless you turn on Include sensitive topics in memory.
Projects
Claude Projects let you group conversations around a single topic, pin reference files (docs, briefs, codebases), and reuse a custom system prompt across every chat in that Project. Use them for ongoing work like a book, a client project, or a research topic you keep coming back to. Claude keeps the context loaded so you don’t paste the same files every time.
Skills
Claude Skills are small reusable instruction packs Claude loads only when relevant. Install a Skill once, and Claude knows how to do that task, whether it’s formatting an internal report a specific way, running a recurring workflow, or applying a brand voice. Anthropic ships some, the community ships some, and you can build your own.
Connectors and MCP
Connectors plug Claude into the tools you already use (Gmail, Google Drive, Calendar, Notion, GitHub, Slack, and more) so it can read, search, and, with your permission, take actions inside those services. Underneath, this runs on the Model Context Protocol (MCP), the open standard Anthropic introduced so any tool can speak to Claude (and other AI systems) the same way.
Cowork
Claude Cowork is the part of Claude that handles bigger jobs: research a topic, build a spreadsheet, draft a report or a presentation, and schedule tasks to run later. Since September 2026 it’s built into every conversation on Pro and Max (web, desktop and mobile), so you don’t pick a mode first. Ask a quick question or hand over a bigger task in the same chat.
Artifacts
Artifacts are interactive outputs Claude builds in a side panel: working web apps, dashboards, charts, slide decks, and documents you can edit and share. You can use them to build a simple working tool, like a budget tracker or a quiz, without writing code.
Claude in Chrome
Pro and higher plans include Claude in Chrome, which lets Claude work inside your web browser. They also include Claude for Microsoft 365. See our Claude in Chrome guide for setup and safe-use tips.
Get smarter with Claude every day
Claude ships new features almost every month. We track each one in plain English.
The daily Beginners in AI newsletter sends you one short tip a day on getting more from Claude, ChatGPT, Gemini, and the rest. Free, every morning.
Get the daily newsletter →Free vs Paid: Should You Upgrade?
Claude’s free plan covers a lot: chat on web, desktop and mobile, web search, voice mode, memory, connectors, Skills, Claude Design, Slides and Docs, and up to 5 Projects, using the Sonnet and Haiku models with daily usage limits. If you ask a few questions a day, free is fine.
Claude Pro is $20 a month, or $17 a month billed annually ($200 up front). Pro adds more usage, Opus 5.5, unlimited Projects, Research mode, Claude Code, Claude in Chrome, the ability to hand off and schedule tasks, and access to the Claude Duet waitlist in the US. It also includes two beta extras: Claude Dashboards, and Claude Science, a separate desktop app for scientific research. Fable 5.1 is available on Pro through pay-as-you-go usage credits. If you use Claude several times a day, Pro is usually worth it.
For heavy users there’s Max, from $100 a month, with 5x or 20x Pro’s usage. Max includes Fable 5.1, which can use up to half of your weekly limit, and $100 to $200 a month in credits for building on the Claude Platform. For companies, Team and Enterprise plans add admin controls and separate business data terms. Current prices are on Anthropic’s pricing page.
Want help putting this into practice?
Join the free Beginners in AI community on Skool. You get a 14-day starter path with one small build a day, plus a place to ask questions when you get stuck. No experience needed.
Join the free community →Free to joinHow do you get better results from Claude?
The single highest-leverage skill for using Claude well is prompting. Claude is more forgiving than most AI tools, but it still performs dramatically better with structured input. Our guide on how to write AI prompts that actually work walks through the 4-part formula (Role, Task, Context, Constraints) with 10 before-and-after examples.
For Claude specifically, three tips that consistently improve results:
- Tell Claude what you don’t want. “Don’t use the phrase ‘delve into.’” “Avoid corporate buzzwords.” Negative instructions shape output almost as much as positive ones.
- Give it examples. If you want your writing in a specific style, paste three samples and say “match this voice.”
- Ask for revisions specifically. “Make it 30% shorter and remove the intro” works better than “make it better.”
How does Claude handle privacy and safety?
Anthropic’s approach to AI safety is one of the reasons Claude exists. The company was founded specifically around making AI more trustworthy. In practice, this means Claude is more likely to refuse harmful requests, but also more willing to engage thoughtfully with nuanced topics than some competitors.
On data privacy: on the consumer plans (Free, Pro and Max), model training is opt-out. Anthropic can use your chats to improve Claude while the Model Improvement setting is on, so check your Privacy Settings and switch it off if you’d rather it didn’t. Incognito chats are never used for training, even with that setting on. Business plans (Team, Enterprise and the API) run under separate commercial terms. Full details are in Anthropic’s model-training FAQ and privacy policy.
Your Next Steps
The fastest way to learn Claude is to use it daily for a week on real tasks. Pick one thing you’d normally do the hard way (drafting an email, summarizing a document, brainstorming) and hand it to Claude instead. After a week you’ll know which jobs it handles well and which it doesn’t.
- Sign up at claude.ai
- Try the 4-part prompt formula on your first real task
- Open Settings and look at Memory and Privacy, so you decide what Claude keeps and whether your chats help train it
- Start a Project for one piece of work you come back to every week
- Take a free course at Anthropic Academy
- Compare Claude to other options in our ChatGPT vs Claude vs Gemini breakdown
If you want ongoing help making the most of Claude and other AI tools, our free daily newsletter covers practical AI tutorials for non-technical professionals. Each one takes about three minutes to read.
How is Claude used in specific roles and industries?
If you have a specific role or industry, here are practical guides for using Claude in that work: what it does well, what it doesn’t, and the workflows that pay off fastest.
By role or business
Made for regular people, by a real person
New Claude features and what they mean for you, plus tips that work with any AI tool. One free email a day. Curated and edited by a human, in plain English.
Free forever. Unsubscribe anytime.
How does Claude compare to ChatGPT and Gemini at a glance?
| Capability | Claude | ChatGPT | Gemini |
|---|---|---|---|
| Best for agentic coding | Yes (Claude Code) | Codex (app, CLI, cloud) | Antigravity, Jules |
| Best for live web search | Good (every plan) | Yes (Search) | Strong |
| Best for Google Workspace | Connectors | Plugins (Gmail, Drive) | Yes (native) |
| Best for long-context (500+ pages) | Yes (up to 1M) | Good (256K on Plus) | Yes (1M on AI Pro) |
| Best free tier | Good (search, voice, memory) | Good | Good (Flash-Lite) |
| Starting paid price | $17/mo (Pro, annual) | $8/mo (Go) | $19.99/mo (AI Pro) |
| Image generation | No (designs and charts only) | Native (Images 2.5) | Native (Nano Banana) |
| Voice mode | Every plan | Strong (ChatGPT Voice) | Strong (Gemini Live) |
For the deeper three-way head-to-head with current pricing and capability scores, see our ChatGPT vs Claude vs Gemini comparison.
What non-obvious Claude tips actually save time?
- Use Projects for ongoing work. A Project in claude.ai holds your files (PDFs, transcripts, notes) and a set of instructions that load in every chat inside it, so you stop re-uploading the same files and re-explaining what you want.
- Upload the real thing. Instead of describing a document, a screenshot or a spreadsheet, attach it. Claude reads PDFs, images and files directly, and its answers get much more specific.
- Check anything that matters. Claude can be confidently wrong. For numbers, dates, medical, legal or money questions, ask it to show its sources or turn on web search, and check the key facts yourself.
- Ask Claude to plan first, then write. “Plan your answer in 3 bullet points first; then write the full answer.” This one habit makes longer answers noticeably better.
- Use Artifacts for anything you’ll keep or reuse. Claude opens it in a side panel (an interactive document, code or a diagram) instead of pasting it into the chat. See our Claude Artifacts guide for details.
- Switch models mid-conversation. Use Sonnet 5.5 or Haiku 5.5 for routine work and move up to Opus 5.5 only when the task needs deeper reasoning. People who stay on the biggest model by default burn through their limits faster than they need to.
- Check what Claude remembers. Memory is on by default, so open Settings > Memory now and then to correct or delete anything out of date.
- Try the desktop app for work on your own files. The Claude desktop app reads local files and supports MCP servers and desktop extensions. Long tasks now run in any chat on web, desktop or mobile, so the desktop app matters most when Claude needs your computer’s files. See our Claude Desktop App guide.
- Let Claude in Chrome do repetitive web tasks. On Pro and above, Claude in Chrome can work in your browser for you. It’s best for steps you repeat often and can describe one at a time. Our Claude in Chrome guide walks through setup.
- Know what happens at your limit. Every plan has usage limits. When you hit one, Claude tells you when it resets. Shorter chats and lighter models make it last longer; our usage limits guide has more ways to stretch it.
Frequently asked questions about using Claude
Is Claude free?
Yes. The free plan at claude.ai includes Sonnet and Haiku with daily usage limits, plus web search, voice mode and memory. Heavier use requires Claude Pro ($17/month billed annually or $20 monthly), Max, Team, or Enterprise. See our full Claude pricing breakdown.
What’s the difference between Claude.ai and Claude API?
claude.ai is the consumer product (chat interface, free or subscription). The Claude Platform is the developer/API product (pay-as-you-go, programmatic access). Most users only need claude.ai.
Should I use Opus, Sonnet, or Haiku?
Default to Sonnet 5.5 for most work. Switch to Opus 5.5 (Pro and above) for hard reasoning, multi-step tasks, or long documents, and to Haiku 5.5 for quick, high-volume tasks. Fable 5.1 sits above Opus for the very biggest jobs and is on paid plans only.
Does Claude have memory?
Yes. Claude remembers details across your chats on every plan, and memory is on by default for Free, Pro and Max. You can review, edit or delete what it remembers in Settings > Memory. Projects are still the best place to keep files and instructions for one piece of work together.
Can Claude access the internet?
Yes, even on the free plan. Claude searches the web when a question needs current information. For longer reports built from many sources, Pro and above add Research mode.
Does Anthropic train on my Claude conversations?
On Free, Pro and Max, Anthropic can use your chats to improve its models while the Model Improvement setting is on. You can turn this off in your Privacy Settings. Incognito chats are never used for training. Team and Enterprise accounts run under separate business terms. See Anthropic’s model-training FAQ.
How is Claude different from ChatGPT?
Different strengths: Claude leads on writing quality, long-context reasoning, and agentic coding. ChatGPT leads on image generation and has more ready-made integrations. See our full ChatGPT vs Claude comparison.
Can I use Claude in my browser without an account?
You need a free Anthropic account to use claude.ai. Sign-up is free; usage on the free tier has daily caps.
What is Claude Mythos?
Claude Mythos 5.1 is Anthropic’s restricted model, released alongside Fable 5.1 on September 1, 2026. It’s available only to organizations verified through Anthropic’s trusted-access programs, not on any personal plan, so Pro and Max subscribers use Fable 5.1 instead. See Anthropic’s announcement.
What is Claude Cowork?
The part of Claude that handles longer jobs such as reports, spreadsheets, presentations and scheduled tasks. On Pro and Max you don’t need to open it separately: just ask in any chat. Read our Claude Cowork guide.
What is Claude Duet?
A separate app, launched October 21, 2026: a personal-assistant version of Claude that messages you first. It sends a morning brief, checks your connected apps and takes on tasks, and it asks before it sends, pays, signs or deletes anything. It’s a research preview for Pro and Max users in the US, through a waitlist on the Claude Duet page.
What is Claude Code?
Anthropic’s AI coding assistant for developers. It runs in your terminal, a code editor, the desktop app or the browser, and reads your real project. It needs a paid plan. Read our Claude Code guide.
Sources and official Anthropic documentation
- Claude, the consumer product
- Anthropic news: official announcements and release notes
- Anthropic developer documentation
- Claude models overview: capabilities and pricing
- Claude pricing: current tiers
- Claude Platform: API access and billing
- Anthropic privacy policy
- Introducing Claude Opus 5.5 and Claude Sonnet 5.5
- Introducing Claude Duet and Get started with Claude Duet
- Is my data used for model training?
- Claude app release notes (memory, Cowork, web search)
For concept primers, see our glossary entries on AI tokens explained, the broader AI glossary hub, and our coverage of the Anthropic Academy for free official Anthropic courses.
You May Also Like
- Where to Start with Claude: Your First 30 Minutes
- How to Write AI Prompts That Actually Work
- Best Claude Prompts: 50 Examples
- Claude Pricing: Free vs Pro vs Max
- Claude Usage Limits: How They Work and How to Stretch Them
- What Claude Remembers About You
- Claude Desktop App: Everything You Need to Know
- Skills vs Connectors vs Plugins: How to Extend Claude
- Claude Opus 5.5: What Changed
- Anthropic Academy: Free Claude Courses
- How to Get Honest Pushback From Claude
- Is Claude Down? How to Check