How to Set Up Claude

Go to claude.ai and click Sign Up. Register with an email address or sign in with your existing Google or Apple account. You will verify your email and complete a quick phone verification, then land directly in the chat interface.

The free tier gives you access to Claude Sonnet 4.6 with web search, memory, file uploads, and Artifacts. That is enough to evaluate whether Claude fits your work before spending anything. No credit card is required.

Claude offers multiple models at different capability levels. Sonnet handles everyday tasks fast. Opus provides deeper reasoning for complex analysis, longer documents, and harder coding problems but requires a paid plan. You do not need to choose yet. Start with the free tier and upgrade when you consistently want more depth or hit the usage ceiling.

Start free. Most people know within two to three days of regular use whether the free tier covers their needs.

If you upgrade, Pro at $20 per month is the right choice for most individuals. It unlocks Opus, Projects for persistent workspaces, extended thinking for complex reasoning, Deep Research, and roughly five times the free tier’s usage allowance. Usage runs on a rolling five hour window, not a daily reset, so spacing sessions out is more effective than cramming everything into one sitting.

Move to Max ($100 or $200 per month) only if you use Claude for the majority of your workday and regularly hit Pro limits during focused sessions. The two Max tiers offer five times and twenty times Pro capacity.

For organizations, Team Standard at $30 per seat per month ($25 billed annually) adds shared workspaces, admin controls, SAML SSO, and centralized billing with a five seat minimum. Enterprise plans are custom priced with data residency, audit logging, and compliance certifications.

Do this before your first real work conversation, not after. Click your profile icon in the bottom left, select Settings, then review the privacy controls.

On the free and Pro tiers, Anthropic may use your conversations to improve its models. Toggle this off. If you handle anything proprietary or client related, do it before uploading a single document. Team and Enterprise plans exclude conversations from training by default.

For regulated industries, Enterprise plans add data residency controls, audit logging, and compliance certifications that satisfy most procurement security reviews. The setting change takes thirty seconds. Cleaning up after you have already uploaded confidential data to an account that was still opted into training is a much longer conversation with your security team.

Claude works in your browser, but the desktop app removes the friction of finding and switching to a browser tab. Download it for macOS or Windows from claude.ai/download.

The desktop app adds system wide keyboard shortcuts that summon Claude from any application, voice mode for hands free interaction, and direct file drag and drop from your file system. On paid plans, it also enables Cowork, which lets Claude interact with files and applications on your computer to complete multistep tasks autonomously.

There is also a Chrome extension (Claude in Chrome, currently in beta) that lets Claude assist with browsing and web research directly in your browser. If you use Claude more than a few times per day, the desktop app is worth installing on day one.

Claude is available on iOS and Android through the App Store and Google Play. Sign in with your existing account. Conversations, Projects, memory, and all plan features sync across devices.

The mobile app adds voice mode and camera input. Point your camera at a whiteboard sketch, a printed contract, or a restaurant receipt, and Claude can analyze it in the conversation. This is useful for capturing ideas on the go, getting quick answers about physical documents, or continuing a work conversation during a commute.

Anything started on mobile picks up exactly where you left off on desktop. The sync is instant.

This is the most important configuration step in the entire setup. In the left sidebar, click Projects, then Create a project. Name it after your primary work context: “Q3 Product Launch,” “Client Proposals,” or “Engineering: Auth Service.”

In the project instructions field, write the context that should apply to every conversation. You get roughly 8,000 characters here, which is plenty for a detailed brief. Be specific: “I am a product manager at a Series B SaaS company. Our sprints run two weeks. I prefer concise responses with bullet points for action items. Our tech stack is React, Node, and PostgreSQL.” Vague instructions like “be helpful” produce almost nothing useful because they do not change how Claude actually responds.

Upload two to three reference documents (a project brief, style guide, or technical spec). Every conversation inside this Project draws on those instructions and files automatically, so you stop pasting the same context into every chat.

Memory lets Claude retain facts and preferences across all conversations without you repeating them. Go to Settings > Capabilities and make sure both memory toggles are on: “Search and reference chats” (lets Claude look back at your history) and “Generate memory from chat history” (lets Claude learn from conversations over time).

Then tell Claude things explicitly. “Our fiscal year starts in April” or “I prefer tables over bullet lists for data comparisons.” Claude picks up preferences faster from direct statements than from behavioral patterns over time.

Your profile preferences (under Settings > Profile) have a 1,500 character limit, and every character consumes tokens in every conversation. Keep them tight: your role, your formatting preferences, and the five to eight rules that matter most. Anything project specific belongs in a Project, not here. You can also pick a default writing style under Settings > Styles: choose Concise, Formal, Explanatory, or create a custom style.

MCP (Model Context Protocol) connects Claude to your external tools and data sources. In the chat interface, open the connectors menu (the tools icon below the message input). Available connectors include Google Drive, Gmail, Google Calendar, Slack, GitHub, Asana, Jira, and Figma, with new ones added regularly.

Free users can connect one custom connector at a time. Pro and above have no connector limit. Start with the one tool where your most important context lives. Once connected, Claude can pull data directly into conversations: “summarize my unread emails from this morning,” “find the Q3 budget spreadsheet in Drive,” or “what meetings do I have tomorrow.”

MCP is strongest for retrieval and analysis. Claude can search your files, surface email threads, and check your calendar, but it cannot send emails or create calendar events through most connectors. The value is giving Claude context about your actual work so its responses draw on real data. If a connector fails to load, disable your browser extensions first. Ad blockers and privacy tools are the most common culprit.

Claude Code is an agentic coding tool that reads your entire codebase, writes and edits files, runs terminal commands, creates tests, and commits to git. It is available as a terminal CLI, a VS Code extension, a JetBrains plugin, and through the web interface at claude.ai/code.

Install the CLI with a single command: curl -fsSL https://claude.ai/install.sh | bash on macOS and Linux, or irm https://claude.ai/install.ps1 | iex on Windows PowerShell. Then run claude in your terminal and authenticate through your browser.

Start with a small project. Ask Claude Code to explain the codebase structure, then try a focused task: a refactor, a test suite for an untested module, or a bug fix. One thing to know: Claude Code and Claude.ai chat share the same usage quota, so a heavy coding session in the morning can leave you with reduced chat capacity in the afternoon.