Claude Code FAQ: 15 Common Questions, Answered

These are the questions we see most often from people starting with Claude Code, answered in a few lines each, with a link to the deeper guide when there is one. Every answer was checked against the official Claude Code documentation at code.claude.com on October 24, 2026.

Getting started

What is Claude Code?

Claude Code is Anthropic’s agentic coding tool. You describe a task and it reads your codebase, edits files and runs commands, asking permission according to the mode you choose. It runs in the terminal and also has a desktop app, VS Code and JetBrains extensions, and a web version. Start with our getting started guide, then try the hands-on Claude Code tutorial.

How do I install Claude Code?

The recommended native installer:

# macOS, Linux, WSL
curl -fsSL https://claude.ai/install.sh | bash

# Windows PowerShell
irm https://claude.ai/install.ps1 | iex

Homebrew (brew install --cask claude-code), WinGet (winget install Anthropic.ClaudeCode) and npm (npm install -g @anthropic-ai/claude-code, Node.js 22 or later) also work. Then run claude in your project folder and sign in. Verify with claude --version; for diagnostics, claude doctor.

Is Claude Code free to use?

The software is free to install, but you need a paid account to use it: a Claude Pro, Max, Team or Enterprise plan, a Claude Console account billed per token, or a cloud provider such as Amazon Bedrock, Google Cloud’s Agent Platform or Microsoft Foundry. The free Claude.ai plan does not include Claude Code.

How much does it cost?

As of October 24, 2026, Pro is $20/month (or $17/month billed annually) and Max is $100 or $200/month. On the API you pay per token. Full breakdown, including Team seats and API rates, in our Claude Code pricing guide.

Does Claude Code work on Windows?

Yes. Claude Code runs natively on Windows 10 1809 or later and Windows Server 2019 or later, installed from PowerShell or CMD, or inside WSL. Sandboxing is only supported under WSL 2, and Git for Windows is recommended so Claude Code can use the Bash tool. Step-by-step commands for PowerShell, CMD and WSL are in how to install Claude Code.

Can I use it offline?

No. Claude Code runs locally but sends prompts and context to the model over the network, so an internet connection is required.

Models and limits

Which models does Claude Code use, and how do I switch?

Claude models only. Switch mid-session with /model, or at launch with claude --model <alias>, using aliases such as sonnet, opus, haiku or fable. As of October 2026 the default on Pro, Max, Team, Enterprise and the Anthropic API is Opus 5.5. Anthropic’s own advice: Sonnet for most coding, Opus for complex reasoning.

claude --model sonnet
# inside a session
/model opus

Why did I hit a usage limit so fast?

On subscriptions, limits are a five-hour window plus a weekly cap, shared with Claude.ai chat. Long sessions also get more expensive per message, because the whole conversation is re-sent on every request. Run /usage to see what is consuming your allowance, and /clear between unrelated tasks. The context management guide explains /context, /compact and /clear.

Safety and privacy

Does Anthropic train on my code?

It depends on your account. On Free, Pro and Max, your code and prompts are used for training only if you turn on the model-improvement setting. Under commercial terms (Team, Enterprise, API and cloud providers), Anthropic does not train generative models on Claude Code data unless your organisation opts in to a program such as the Development Partner Program.

Retention also differs: commercial accounts have a standard 30-day retention period; consumer accounts get 30 days with model improvement off and 5 years with it on. Details are on the official data usage page.

How do I stop Claude from running commands without asking?

Use permission modes and rules. The default mode asks before edits and commands; plan mode only reads and plans; Shift+Tab cycles modes during a session. Add deny rules in .claude/settings.json for things that must never happen, and a PreToolUse hook for hard guarantees. See the permission modes guide and the hooks guide.

How do I undo changes Claude Code made?

Press Esc twice with an empty prompt, or run /rewind, to open the rewind menu. You can restore code, conversation, or both to any earlier prompt. Checkpoints only cover edits made with Claude’s file editing tools, so keep using Git for anything else.

Two caveats from the docs: files changed only by Bash commands (for example a migration script) aren’t tracked, and edits made by most subagents aren’t restored. Commit before big tasks.

Configuration

What is CLAUDE.md, and do I need one?

CLAUDE.md is a Markdown file of project instructions Claude Code loads at the start of every session: build commands, conventions, “never do X” rules. Run /init to generate a starter file from your codebase, then trim it; Anthropic recommends keeping it under about 200 lines. Our CLAUDE.md guide has templates.

Can Claude Code read AGENTS.md?

Yes, in recent versions. Claude Code reads AGENTS.md as project instructions when there is no CLAUDE.md or CLAUDE.local.md in your working directory or above it. If you keep both, import the shared file from CLAUDE.md with @AGENTS.md so Claude reads it too.

# CLAUDE.md
@AGENTS.md

## Claude-specific notes
- Use plan mode for changes under src/billing/

How do I connect MCP servers?

Use claude mcp add, or commit a .mcp.json file so your team shares the same servers. For example, a remote HTTP server:

claude mcp add --transport http <name> <url>

Run /mcp inside a session to check status and authenticate. See the MCP setup guide and our list of the best MCP servers for Claude Code.

How do I continue yesterday’s session?

claude -c                 # continue the most recent conversation in this directory
claude -r                 # pick a session from an interactive list
claude -r "auth-refactor" # resume a named session

Name a session with /rename before you leave so it’s easy to find. Session transcripts are stored locally under ~/.claude/projects/ (30 days by default, adjustable with cleanupPeriodDays).

Going further

Once the basics click, the next step is extending Claude Code: skills for repeatable procedures, subagents for isolated work, hooks for guarantees. Our explainer on skills vs subagents vs hooks shows which to use when, the slash commands cheatsheet covers the built-in commands, and what is a Claude kit? shows how the pieces fit into one reusable setup. If you’re still choosing a tool, compare Claude Code vs Codex and Claude Code vs Cursor.