# Claude Code

Published January 1, 2026 · Updated August 10, 2026

AI-powered development environment by Anthropic

Category: Developer Tools · Paid · [Official website](https://claude.ai)

## Install with Homebrew

`brew install --cask claude-code`

## Quick take

Claude Code remains a top coding agent on Mac in August 2026: Sonnet 5 and Opus 5, Pro inclusion, improved May rate limits, and a mature permissioned tool loop. Budget Pro to start and Max if you live in agents all day. Watch shared usage with chat, review every serious diff, and track new rivals (Codex-in-ChatGPT, Muse Code, Cursor) without abandoning a tool that already fits your terminal life. It is not magic. It is a strong junior-senior hybrid that still needs engineering judgement.

Best for: Terminal-centric developers on Mac, Multi-file refactors and debugging with Anthropic models, Teams standardising on Claude seats with Premium/Enterprise, Builders who want CLAUDE.md-driven project memory, People escaping pure autocomplete tools into real agents

## What is Claude Code?

Claude Code is Anthropic's agentic coding tool for the terminal, IDEs (VS Code and JetBrains), desktop surfaces, and browser-oriented flows. You point it at a repository, describe an outcome in natural language, and it reads files, edits code, runs commands, manages git, and iterates on failures (with your approval gates) until the task is done or blocked.

Unlike autocomplete-first tools that only suggest the next line, Claude Code is built for multi-step agency across a codebase. It loads project guidance from CLAUDE.md files, respects tool permission prompts, and can coordinate subagents for parallel work. In August 2026 it rides Anthropic's current model stack: Claude Sonnet 5 as the efficient default for much coding volume, Claude Opus 5 for hard debugging and long-horizon agent work (Max default / strongest common Pro model), with Haiku-class models for cheap utility steps and Fable 5 available on higher-end access when you need peak intelligence.

Billing is not a separate Claude Code SKU for most people. Anthropic includes Claude Code with paid Claude plans: Pro at $20/month ($17/month annual), Max 5× at $100/month, Max 20× at $200/month, Team Premium seats, and Enterprise. Free chat tiers are not a serious Claude Code plan. You can also bring an API key and pay per token (Opus 5 at $5/$25 per million input/output tokens; Sonnet 5 introductory $2/$10 through 31 August 2026, then $3/$15). On 6 May 2026 Anthropic permanently doubled Claude Code five-hour rate limits for Pro/Max/Team/Enterprise seat plans and removed peak-hours limit reductions for Pro and Max, a direct response to developers hitting walls during workdays.

Competitive landscape since May 2026: OpenAI folded Codex into the ChatGPT desktop app (July 2026) on GPT-5.6; Cursor continues with Composer 2.5 and Teams pricing changes; Meta launched Muse Code (5 August 2026) as a terminal agent narrative rival; GitHub Copilot CLI picked up GPT-5.6 models. Claude Code remains a default recommendation for developers who want deep repo agency with Anthropic models, especially on Mac laptops where terminal-first workflows dominate.

Platform notes: Claude Code is first-class on macOS (Apple Silicon and Intel), Linux, and Windows. Pair it with native terminals (Ghostty, iTerm2, Terminal, Warp) or IDE side panels. Companion utilities like claude-devtools (session inspection) and Vibe Notch / Claude Island (notch status) are popular in the Mac community but are third-party, not Anthropic products.

What changed since the May 2026 content corpus: model names moved from older Sonnet/Opus 4.x lines to Sonnet 5 and Opus 5; subscription inclusion is the primary packaging story (not "API-only mysterious pricing"); rate limits improved in May; new terminal rivals appeared (Muse Code); and usage is explicitly shared with Claude chat/Cowork on the same plan, a long session in the desktop app can consume the same pool you wanted for coding.

Operationally, teams adopt Claude Code in three patterns. Solo developers authenticate with Pro or Max and live in the terminal. Agencies issue Team Premium seats so Claude Code is entitled and centrally billed. Enterprises wire SSO, set spend limits, and sometimes prefer API keys with organization policies for CI agents. Hybrid patterns are common: interactive coding on Max, batch refactor bots on API with hard budgets.

Safety culture matters. Claude Code can run shell commands and edit files; treat permissions seriously, use branch protections, and review diffs. Anthropic continues to ship safeguards (including cyber-related classifier behaviours on frontier models) that can fall back models on flagged requests. For regulated codebases, combine Enterprise contractual terms with local secret hygiene, never paste production credentials into prompts when 1Password/MCP-style secret brokering is available.

Mac-specific ergonomics: run Claude Code inside tmux or a dedicated Ghostty profile; keep notifications via Vibe Notch if you multitask; inspect opaque sessions with claude-devtools when the terminal only shows summaries; and keep Xcode/Android Studio for builds while Claude handles repo reasoning. Apple Silicon M-series machines handle large context windows comfortably for Sonnet 5/Opus 5 workloads compared with older Intel laptops with limited RAM, 16GB is workable, 32GB+ is happier for big monorepos.

Extended guidance for monorepo teams: create a root CLAUDE.md that points to package-specific instruction files rather than pasting thousands of lines into one blob. Agents lose signal when instruction files become novels. Prefer layered docs: root for repo navigation, package files for local test commands, and a SECURITY.md section listing paths the agent must never touch (secrets, production terraform without human review, signing keys).

On performance, Apple Silicon machines with 32GB unified memory handle multi-agent Opus 5 sessions more gracefully than 8GB machines that swap under node_modules pressure. If your Mac fans spin whenever the agent runs full test suites, constrain the agent to targeted test commands in CLAUDE.md instead of allowing blanket npm test at repo root. Also consider remote builders: Claude Code on a Mac laptop driving tests on a Linux CI box via SSH can be a better thermal and battery story than compiling everything locally.

Procurement checklist for engineering managers: (1) map who needs Premium vs Standard seats, (2) set usage credit policies before a product launch week, (3) require CLAUDE.md in service templates, (4) ban shared personal Max logins, (5) add a monthly retrospective on agent-caused incidents, (6) compare spend against Cursor and Codex seats quarterly rather than forever-renewing on inertia. Claude Code is excellent, not inevitable, keep the competitive bake-off alive as Muse Code and GPT-5.6 Codex evolve.

Migration note from older content: stop describing Claude Code as API-only mystery pricing. The 2026 reality is subscription-first for humans, API-first for automation, with shared limits and explicit Max tiers. Update internal wiki pages that still say Sonnet 3.7 or Opus 4-only guidance, those strings mis-train both humans and agents.

## Features

- **Repository-scale agency.** Claude Code reads project structure, follows imports, and applies multi-file edits with test/command loops. It is optimised for tasks like "add feature X with tests" or "find and fix flaky suite" rather than single-line autocomplete. On Opus 5 it shows stronger root-cause persistence; on Sonnet 5 it covers more ground per dollar during bulk refactors.
- **CLAUDE.md and project memory.** Checked-in instructions teach coding standards, architecture boundaries, and forbidden paths. Combined with per-project memory under ~/.claude/, the agent stops relearning that you use pnpm, bun, or a particular error format. Treat CLAUDE.md as real documentation, review it in PRs.
- **Permissioned tool use.** File reads/writes, bash, web, and MCP tools surface approvals. Power users tune allowlists for trusted commands while keeping destructive operations gated. Companion notch apps can surface approve/deny without alt-tabbing on Mac.
- **IDE and terminal surfaces.** Work from CLI for maximum control or from VS Code/JetBrains integrations when you want inline diffs. The same account entitlements apply. Desktop Claude Cowork is adjacent but not a full replacement for repo-native Claude Code sessions.
- **Subagents and team coordination.** Complex tasks can fan out to subagents with isolated traces. Session inspectors (claude-devtools) make those trees reviewable after the fact when the terminal collapses them.
- **Shared plan limits with chat/Cowork.** Subscription users share five-hour and weekly pools across Claude surfaces. Heavy chat burns coding headroom. Max tiers exist primarily to buy more of that shared pool, not a different Claude Code binary.
- **API and usage credits escape hatches.** When subscription caps hit, paid users can continue via usage credits at API rates or switch fully to Console API keys with org spend caps, critical for deadline weeks after May 2026 limit changes.
- **Model choice: Sonnet 5 vs Opus 5.** Pick Sonnet 5 for volume and cost control (watch tokenizer inflation vs older Sonnets). Pick Opus 5 for subtle bugs, multi-hour agents, and tasks where near-Fable quality matters at Opus price. Fast mode variants trade money for latency on Opus-class work.

## How to Install Claude Code on Mac

Official installs target macOS terminals first; IDE extensions are optional companions.

1. **Get a paid Claude plan or API key.** Subscribe at claude.com (Pro/Max/Team/Enterprise) or create a Console API key with billing limits.
2. **Install Claude Code.** Follow Anthropic's current Claude Code install docs (native installer or package channel). Ensure Node/native dependencies match the doc version you install.
3. **Authenticate.** Sign in with your Claude account or configure the API key environment variables as documented.
4. **Open a repo and add CLAUDE.md.** cd into a git project, create CLAUDE.md with build/test commands and style rules, then start a session with a small task to verify permissions.
5. **Optional IDE + companions.** Install VS Code/JetBrains integrations; consider claude-devtools for session review and Vibe Notch for status on notched MacBooks.

## Pros

- Best-in-class agentic coding quality on Opus 5 / Sonnet 5 for many Mac developers in summer 2026
- Deep repo context and CLAUDE.md configurability beat generic chat paste workflows
- Included in Pro, no separate "coding SKU" required to start
- May 2026 rate limit doubling reduced midday lockouts
- Works in terminal and IDEs without forcing a fork editor like some AI IDEs
- Strong ecosystem of Mac companions (devtools, notch status apps)
- Enterprise and Team admin paths for companies
- API escape hatch when subscription caps bind

## Cons

- Shared limits with chat/Cowork surprise people who also live in Claude desktop
- Pro still not enough for all-day multi-agent work. Max money adds up
- Can make large incorrect edits if you rubber-stamp permissions
- Tokenizer/pricing changes (Sonnet 5) complicate cost prediction vs older Sonnets
- Not a full IDE, you still need Xcode/VS Code/JetBrains for many builds
- New rivals (Muse Code, Codex-in-ChatGPT, Cursor) mean the "only choice" era is over

## Closer look: Running Claude Code as production infrastructure

Claude Code is no longer a toy demo. Teams treat it with the same operational seriousness they give CI. If your internal docs still frame pricing as opaque API metering only, rewrite them: human seats are subscription-first in 2026, automation stays API-first, limits are shared, and Max is an explicit tier rather than a rumour. Also scrub stale model strings such as Sonnet 3.7 or Opus 4-only playbooks; outdated model names mis-train both new hires and the agents that read your wiki.

## FAQ

### How much does Claude Code cost in 2026?

Most individuals pay via Claude Pro ($20/mo or $17 annual) or Max ($100/$200). Teams use Premium seats (~$100-125). Alternatively pay API token rates for Opus 5, Sonnet 5, etc. There is no meaningful long-term free Claude Code tier.

### Is Claude Code included with Claude Pro?

Yes. Pro, Max, Team Premium, and Enterprise include Claude Code subject to shared usage limits. Confirm Team Standard vs Premium seat types before buying for engineers.

### What model should I use. Sonnet 5 or Opus 5?

Default Sonnet 5 for volume and cost. Switch to Opus 5 for deep debugging, multi-hour agents, and high-stakes refactors. Use higher effort settings only when quality gains pay for tokens.

### Did rate limits improve in 2026?

Yes. On 6 May 2026 Anthropic doubled Claude Code five-hour rate limits for Pro/Max/Team/Enterprise seat plans and removed peak-hours reductions for Pro and Max.

### Does Claude Code work on Apple Silicon Macs?

Yes. It is commonly used on M-series MacBooks with modern terminals. Ensure adequate RAM for large monorepos and node toolchains.

### How is Claude Code different from Cursor?

Claude Code is agent-first in terminal/IDE integrations using Anthropic models and CLAUDE.md. Cursor is an AI editor product with its own models and UX. Many developers use both for different tasks.

### What about Meta Muse Code?

Muse Code launched 5 August 2026 as a new terminal coding agent competitor. It is early relative to Claude Code's maturity; evaluate on your repos rather than launch blogs alone.

### Why did my limits disappear after using Claude chat?

Usage is shared across Claude surfaces on subscription plans. Desktop chat, mobile, Cowork, and Claude Code draw one pool.

### Can I use Claude Code offline?

No meaningful offline mode, models and much tooling require network access to Anthropic (or approved VPC deployments for enterprise variants).

### How do I reduce surprise cost?

Prefer Sonnet 5, tighten CLAUDE.md to avoid exploratory thrash, set API spend caps, enable only needed MCP tools, and inspect sessions for wasted tool loops.

### Can Claude Code edit multiple git worktrees?

Yes in practice, open separate sessions per worktree directory. Isolating experiments this way is safer than letting one session hop branches underneath you.

### Should I disable bash tools?

For read-only audits, yes. For build/test loops, no, but allowlist package managers and test runners carefully.

### How does Claude Code relate to Claude Science?

Claude Science is a separate scientific workbench product line launched mid-2026. Some research coding overlaps, but Science is not a substitute for repo-oriented Claude Code on product engineering.

### What terminal is best on Mac?

Any modern terminal works. Ghostty, iTerm2, Warp, and Apple Terminal are all common. Pick based on performance and font rendering; Claude Code is not terminal-exclusive in capability.

### Can I run Claude Code in CI?

Yes via API keys and noninteractive patterns, with strict spend caps. Prefer ephemeral runners and never bake long-lived org keys into public forks.

### Can Claude Code use GPT models?

Claude Code is built around Anthropic Claude models (Sonnet 5, Opus 5, etc.). If you need GPT-5.6 agents, use Codex-in-ChatGPT, Copilot, or Cursor with OpenAI models instead of expecting model multipath inside Claude Code.

### How do I handle secrets?

Keep secrets in a password manager or secret manager; reference them via environment injection rather than pasting into prompts. Deny tools that cat credential files unless absolutely required, and audit with claude-devtools triggers.

### Does Claude Code support Swift and Xcode projects?

Yes, it edits Swift/source files and can run shell build/test commands you allow. You still use Xcode for GUI debugging, Instruments, and signing workflows the agent should not own unsupervised.

## Sources

- [Claude Code product page](https://www.anthropic.com/claude-code)
- [Claude Pricing](https://claude.com/pricing)
- [Introducing Claude Opus 5](https://www.anthropic.com/news/claude-opus-5)
- [Introducing Claude Sonnet 5](https://www.anthropic.com/news/claude-sonnet-5)
- [Higher limits announcement](https://www.anthropic.com/news/higher-limits-spacex)
- [Claude Code pricing guide 2026](https://www.superblocks.com/blog/claude-code-pricing)
- [Usage limits timeline 2026](https://explainx.ai/blog/claude-usage-limits-2026-timeline-explained)
- [Meta Muse Code launch coverage](https://9to5mac.com/2026/08/05/meta-launches-muse-code-ai-coding-agent-for-macos-and-linux/)

## Related

- [Cursor](https://bundl.run/apps/cursor)
- [ChatGPT](https://bundl.run/apps/chatgpt)
- [Claude](https://bundl.run/apps/claude)
- [Codex](https://bundl.run/apps/codex)
- [Windsurf](https://bundl.run/apps/windsurf)
- [Kiro](https://bundl.run/apps/kiro)
- [Claude Code vs Codex](https://bundl.run/compare/claude-code-vs-codex)
- [Claude Code vs GitHub Copilot CLI](https://bundl.run/compare/claude-code-vs-copilot-cli)
- [Claude Code vs OpenCode](https://bundl.run/compare/opencode-desktop-vs-claude-code)

```json
{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "Organization",
      "@id": "https://bundl.run/#organization",
      "name": "Bundl.run",
      "url": "https://bundl.run",
      "logo": {
        "@type": "ImageObject",
        "url": "https://bundl.run/og-image.png",
        "width": 1200,
        "height": 630
      },
      "description": "The Ninite for Mac. Install all your essential Mac apps with one terminal command.",
      "sameAs": [
        "https://github.com/abhiofficial/bundl-mac-setup",
        "https://x.com/bundlrun",
        "https://www.producthunt.com/products/bundl-run"
      ],
      "foundingDate": "2024",
      "contactPoint": {
        "@type": "ContactPoint",
        "contactType": "customer support",
        "url": "https://bundl.run/faq"
      }
    },
    {
      "@type": "WebSite",
      "@id": "https://bundl.run/#website",
      "name": "Bundl.run",
      "url": "https://bundl.run",
      "description": "The Ninite for Mac. Install all your essential Mac apps with one terminal command.",
      "publisher": {
        "@id": "https://bundl.run/#organization"
      },
      "inLanguage": "en-US"
    },
    {
      "@type": "Person",
      "@id": "https://bundl.run/authors/alex-chen#person",
      "name": "Alex Chen",
      "jobTitle": "Senior Developer Tools Specialist",
      "url": "https://bundl.run/authors/alex-chen",
      "worksFor": {
        "@id": "https://bundl.run/#organization"
      },
      "description": "Alex Chen has been evaluating developer tools and productivity software for over 12 years, with deep expertise in code editors, terminal emulators, and development environments. As a former software engineer at several Bay Area startups, Alex brings hands-on experience with the real-world workflows these tools are meant to enhance. Alex tests each application extensively on both Intel and Apple Silicon Macs, documenting performance metrics, integration capabilities, and workflow efficiency. When not reviewing software, Alex contributes to open-source projects and writes technical tutorials for the developer community.",
      "knowsAbout": [
        "Code Editors & IDEs",
        "Terminal Emulators",
        "Version Control Tools",
        "DevOps & CI/CD",
        "API Development",
        "Performance Benchmarking"
      ],
      "image": "https://bundl.run/authors/alex-chen.svg"
    },
    {
      "@type": "BreadcrumbList",
      "@id": "https://bundl.run/apps/claude-code#breadcrumb",
      "itemListElement": [
        {
          "@type": "ListItem",
          "position": 1,
          "name": "Home",
          "item": "https://bundl.run"
        },
        {
          "@type": "ListItem",
          "position": 2,
          "name": "Apps",
          "item": "https://bundl.run/apps"
        },
        {
          "@type": "ListItem",
          "position": 3,
          "name": "Claude Code",
          "item": "https://bundl.run/apps/claude-code"
        }
      ]
    },
    {
      "@type": "WebPage",
      "@id": "https://bundl.run/apps/claude-code",
      "url": "https://bundl.run/apps/claude-code",
      "name": "Claude Code for Mac",
      "description": "AI-powered development environment by Anthropic",
      "isPartOf": {
        "@id": "https://bundl.run/#website"
      },
      "publisher": {
        "@id": "https://bundl.run/#organization"
      },
      "inLanguage": "en-US",
      "datePublished": "2026-01-01T00:00:00Z",
      "dateModified": "2026-08-10T13:34:04.000Z",
      "author": {
        "@id": "https://bundl.run/authors/alex-chen#person"
      },
      "mainEntity": {
        "@id": "https://bundl.run/apps/claude-code#software"
      },
      "breadcrumb": {
        "@id": "https://bundl.run/apps/claude-code#breadcrumb"
      }
    },
    {
      "@type": "SoftwareApplication",
      "@id": "https://bundl.run/apps/claude-code#software",
      "name": "Claude Code",
      "description": "AI-powered development environment by Anthropic",
      "applicationCategory": "DeveloperApplication",
      "operatingSystem": "macOS",
      "url": "https://bundl.run/apps/claude-code",
      "image": {
        "@type": "ImageObject",
        "url": "https://img.logo.dev/claude.ai",
        "caption": "Claude Code app icon for Mac"
      },
      "sameAs": [
        "https://claude.ai/",
        "https://formulae.brew.sh/cask/claude-code"
      ],
      "offers": {
        "@type": "Offer",
        "price": "20",
        "priceCurrency": "USD",
        "availability": "https://schema.org/InStock"
      },
      "review": {
        "@type": "Review",
        "author": {
          "@id": "https://bundl.run/authors/alex-chen#person"
        },
        "reviewBody": "Claude Code remains a top coding agent on Mac in August 2026: Sonnet 5 and Opus 5, Pro inclusion, improved May rate limits, and a mature permissioned tool loop. Budget Pro to start and Max if you live in agents all day. Watch shared usage with chat, review every serious diff, and track new rivals (Codex-in-ChatGPT, Muse Code, Cursor) without abandoning a tool that already fits your terminal life. It is not magic. It is a strong junior-senior hybrid that still needs engineering judgement."
      },
      "about": {
        "@type": "Thing",
        "name": "Claude Code",
        "description": "AI-powered development environment by Anthropic"
      },
      "isPartOf": {
        "@id": "https://bundl.run/#website"
      }
    },
    {
      "@type": "Article",
      "@id": "https://bundl.run/apps/claude-code#article",
      "headline": "Claude Code for Mac — Full Review & Installation Guide 2026",
      "description": "Claude Code remains a top coding agent on Mac in August 2026: Sonnet 5 and Opus 5, Pro inclusion, improved May rate limits, and a mature permissioned tool loop. Budget Pro to start and Max if you live in agents all day. Watch shared usage with chat, review every serious diff, and track new rivals (Codex-in-ChatGPT, Muse Code, Cursor) without abandoning a tool that already fits your terminal life. It is not magic. It is a strong junior-senior hybrid that still needs engineering judgement.",
      "image": "https://img.logo.dev/claude.ai",
      "author": {
        "@id": "https://bundl.run/authors/alex-chen#person"
      },
      "publisher": {
        "@id": "https://bundl.run/#organization"
      },
      "datePublished": "2026-01-01T00:00:00Z",
      "dateModified": "2026-08-10T13:34:04.000Z",
      "mainEntityOfPage": {
        "@type": "WebPage",
        "@id": "https://bundl.run/apps/claude-code"
      },
      "articleSection": "DeveloperApplication",
      "speakable": {
        "@type": "SpeakableSpecification",
        "cssSelector": [
          "h1",
          ".key-facts"
        ]
      },
      "about": {
        "@type": "SoftwareApplication",
        "name": "Claude Code",
        "url": "https://bundl.run/apps/claude-code"
      }
    },
    {
      "@type": "FAQPage",
      "@id": "https://bundl.run/apps/claude-code#faq",
      "mainEntity": [
        {
          "@type": "Question",
          "name": "How much does Claude Code cost in 2026?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Most individuals pay via Claude Pro ($20/mo or $17 annual) or Max ($100/$200). Teams use Premium seats (~$100-125). Alternatively pay API token rates for Opus 5, Sonnet 5, etc. There is no meaningful long-term free Claude Code tier."
          }
        },
        {
          "@type": "Question",
          "name": "Is Claude Code included with Claude Pro?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Yes. Pro, Max, Team Premium, and Enterprise include Claude Code subject to shared usage limits. Confirm Team Standard vs Premium seat types before buying for engineers."
          }
        },
        {
          "@type": "Question",
          "name": "What model should I use. Sonnet 5 or Opus 5?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Default Sonnet 5 for volume and cost. Switch to Opus 5 for deep debugging, multi-hour agents, and high-stakes refactors. Use higher effort settings only when quality gains pay for tokens."
          }
        },
        {
          "@type": "Question",
          "name": "Did rate limits improve in 2026?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Yes. On 6 May 2026 Anthropic doubled Claude Code five-hour rate limits for Pro/Max/Team/Enterprise seat plans and removed peak-hours reductions for Pro and Max."
          }
        },
        {
          "@type": "Question",
          "name": "Does Claude Code work on Apple Silicon Macs?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Yes. It is commonly used on M-series MacBooks with modern terminals. Ensure adequate RAM for large monorepos and node toolchains."
          }
        },
        {
          "@type": "Question",
          "name": "How is Claude Code different from Cursor?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Claude Code is agent-first in terminal/IDE integrations using Anthropic models and CLAUDE.md. Cursor is an AI editor product with its own models and UX. Many developers use both for different tasks."
          }
        },
        {
          "@type": "Question",
          "name": "What about Meta Muse Code?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Muse Code launched 5 August 2026 as a new terminal coding agent competitor. It is early relative to Claude Code's maturity; evaluate on your repos rather than launch blogs alone."
          }
        },
        {
          "@type": "Question",
          "name": "Why did my limits disappear after using Claude chat?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Usage is shared across Claude surfaces on subscription plans. Desktop chat, mobile, Cowork, and Claude Code draw one pool."
          }
        },
        {
          "@type": "Question",
          "name": "Can I use Claude Code offline?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "No meaningful offline mode, models and much tooling require network access to Anthropic (or approved VPC deployments for enterprise variants)."
          }
        },
        {
          "@type": "Question",
          "name": "How do I reduce surprise cost?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Prefer Sonnet 5, tighten CLAUDE.md to avoid exploratory thrash, set API spend caps, enable only needed MCP tools, and inspect sessions for wasted tool loops."
          }
        },
        {
          "@type": "Question",
          "name": "Can Claude Code edit multiple git worktrees?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Yes in practice, open separate sessions per worktree directory. Isolating experiments this way is safer than letting one session hop branches underneath you."
          }
        },
        {
          "@type": "Question",
          "name": "Should I disable bash tools?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "For read-only audits, yes. For build/test loops, no, but allowlist package managers and test runners carefully."
          }
        },
        {
          "@type": "Question",
          "name": "How does Claude Code relate to Claude Science?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Claude Science is a separate scientific workbench product line launched mid-2026. Some research coding overlaps, but Science is not a substitute for repo-oriented Claude Code on product engineering."
          }
        },
        {
          "@type": "Question",
          "name": "What terminal is best on Mac?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Any modern terminal works. Ghostty, iTerm2, Warp, and Apple Terminal are all common. Pick based on performance and font rendering; Claude Code is not terminal-exclusive in capability."
          }
        },
        {
          "@type": "Question",
          "name": "Can I run Claude Code in CI?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Yes via API keys and noninteractive patterns, with strict spend caps. Prefer ephemeral runners and never bake long-lived org keys into public forks."
          }
        },
        {
          "@type": "Question",
          "name": "Can Claude Code use GPT models?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Claude Code is built around Anthropic Claude models (Sonnet 5, Opus 5, etc.). If you need GPT-5.6 agents, use Codex-in-ChatGPT, Copilot, or Cursor with OpenAI models instead of expecting model multipath inside Claude Code."
          }
        },
        {
          "@type": "Question",
          "name": "How do I handle secrets?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Keep secrets in a password manager or secret manager; reference them via environment injection rather than pasting into prompts. Deny tools that cat credential files unless absolutely required, and audit with claude-devtools triggers."
          }
        },
        {
          "@type": "Question",
          "name": "Does Claude Code support Swift and Xcode projects?",
          "acceptedAnswer": {
            "@type": "Answer",
            "text": "Yes, it edits Swift/source files and can run shell build/test commands you allow. You still use Xcode for GUI debugging, Instruments, and signing workflows the agent should not own unsupervised."
          }
        }
      ],
      "isPartOf": {
        "@id": "https://bundl.run/apps/claude-code"
      }
    },
    {
      "@type": "ItemList",
      "name": "Claude Code Video Tutorials",
      "description": "Tutorial videos for learning Claude Code on Mac",
      "itemListElement": [
        {
          "@type": "ListItem",
          "position": 1,
          "item": {
            "@type": "VideoObject",
            "name": "Claude Code Tutorial for Beginners",
            "description": "You can vibe code with other tools, or you can build durable software with tools like Claude Code. First, I'll walk you through how to install it on Mac or Windows. Next, you’ll connect Claude Code to",
            "thumbnailUrl": "https://i.ytimg.com/vi/eMZmDH3T2bY/mqdefault.jpg",
            "uploadDate": "2025-12-22T15:26:27Z",
            "duration": "PT14M44S",
            "embedUrl": "https://www.youtube.com/embed/eMZmDH3T2bY",
            "interactionStatistic": {
              "@type": "InteractionCounter",
              "interactionType": "https://schema.org/WatchAction",
              "userInteractionCount": 128732
            }
          }
        },
        {
          "@type": "ListItem",
          "position": 2,
          "item": {
            "@type": "VideoObject",
            "name": "The Only Claude Skills Guide You Need (Beginner to Expert)",
            "description": "Learn everything about Claude Agent Skills in this complete guide! I'll show you what Skills are, how they differ from MCPs, slash commands, and subagents, and how to build your own custom skills from",
            "thumbnailUrl": "https://i.ytimg.com/vi/421T2iWTQio/mqdefault.jpg",
            "uploadDate": "2025-10-24T13:00:10Z",
            "duration": "PT36M11S",
            "embedUrl": "https://www.youtube.com/embed/421T2iWTQio",
            "interactionStatistic": {
              "@type": "InteractionCounter",
              "interactionType": "https://schema.org/WatchAction",
              "userInteractionCount": 37665
            }
          }
        },
        {
          "@type": "ListItem",
          "position": 3,
          "item": {
            "@type": "VideoObject",
            "name": "Master 95% of Claude Code in 36 Mins (as a beginner)",
            "description": "My Tools💻 In this video, I break down everything you need to know to use Claude Code effectively for building automations. I cover the interface and what actually matters, the WAT framework we use to",
            "thumbnailUrl": "https://i.ytimg.com/vi/saggDHHnmtQ/mqdefault.jpg",
            "uploadDate": "2026-01-21T13:11:58Z",
            "duration": "PT36M58S",
            "embedUrl": "https://www.youtube.com/embed/saggDHHnmtQ",
            "interactionStatistic": {
              "@type": "InteractionCounter",
              "interactionType": "https://schema.org/WatchAction",
              "userInteractionCount": 27942
            }
          }
        },
        {
          "@type": "ListItem",
          "position": 4,
          "item": {
            "@type": "VideoObject",
            "name": "How To Install Claude Code On Mac",
            "description": "Description: In this tutorial, you will learn how to install and run the AI Coding Assistant, Claude Code, on a Mac computer. In this tutorial, you'll learn: * System Requirements to run Claude Code o",
            "thumbnailUrl": "https://i.ytimg.com/vi/NRF0HoiTSiI/mqdefault.jpg",
            "uploadDate": "2025-08-20T10:00:44Z",
            "duration": "PT14M38S",
            "embedUrl": "https://www.youtube.com/embed/NRF0HoiTSiI",
            "interactionStatistic": {
              "@type": "InteractionCounter",
              "interactionType": "https://schema.org/WatchAction",
              "userInteractionCount": 17381
            }
          }
        },
        {
          "@type": "ListItem",
          "position": 5,
          "item": {
            "@type": "VideoObject",
            "name": "The only Claude Code guide you'll ever need (Opus 4.5)",
            "description": "Here is your complete beginners guide to Claude Code. This is my full end to end workflow. You'll be a MASTER by the end of the video. Claude Code: Visual Studio Code: Claude Design Skill: Initial pro",
            "thumbnailUrl": "https://i.ytimg.com/vi/UVJXh57MgI0/mqdefault.jpg",
            "uploadDate": "2025-12-16T00:33:02Z",
            "duration": "PT24M45S",
            "embedUrl": "https://www.youtube.com/embed/UVJXh57MgI0",
            "interactionStatistic": {
              "@type": "InteractionCounter",
              "interactionType": "https://schema.org/WatchAction",
              "userInteractionCount": 79271
            }
          }
        }
      ],
      "numberOfItems": 5
    }
  ]
}
```