Asset 20 8 2
Does AI recommend your business? Run the free check →

Join 15,000 business owners, marketers and entrepreneurs. The Sunday newsletter you'll be annoyed only arrives once a week.

Article

Claude Code Desktop: How the Code Tab Works

Claude Code desktop is the Code tab inside the Claude desktop app for Mac, Windows and Linux. It lets Claude work on the code in a folder on your computer and lets you review its changes without using a terminal. You need a Pro, Max, Team or Enterprise plan, and you don't have to install Node.js or the command-line version.

If you've seen developers running Claude Code in a black terminal window and thought it wasn't for you, the desktop app is the friendlier door. Here's what the Code tab does, the setting that matters most, and where it differs from the terminal version, all from Anthropic's own documentation.

How to open it

  1. Download the Claude desktop app for macOS or Windows (there's an ARM64 installer for Windows ARM), or install it with apt on Linux, which is in beta.
  2. Open Claude and sign in with your Anthropic account.
  3. Click the Code tab at the top of the sidebar. If you see an upgrade prompt, you need a paid plan first.
  4. Under the prompt box, pick an environment from the Local group, then click the project folder control and choose Open folder.
  5. Choose a permission mode and a model, type your task and press Enter.

Anthropic suggests starting with a small project you know well. I'd go further: start with a copy of it. If you just need the app itself, my guide to the Claude desktop app covers downloading it on each system.

The permission modes, decoded

This is the setting that decides how much Claude can do without asking. You can change it at any time from the mode selector under the prompt box.

ModeWhat Claude doesBest for
ManualAsks before editing files or running commands, and shows a diff you accept or rejectLearning, or anything important
Accept editsAuto-accepts file edits and simple commands like mkdir, touch and mv, but asks before other terminal commandsRoutine edits you'll review afterwards
PlanMaps out an approach before making changesComplex tasks where you want to agree the plan first
AutoRuns without routine prompts while a background classifier checks risky actions match your requestLonger jobs on supported models
Bypass permissionsRuns without permission prompts, apart from a few protected actionsExperienced users only; on Pro and Max you switch it on in Settings

Anthropic notes that earlier versions of the Code tab called these Ask permissions, Auto accept edits and Plan mode, so older tutorials may use those names. Enterprise admins can restrict which modes their team sees.

What the Code tab adds over the terminal

  • Several sessions running in parallel, listed in the sidebar.
  • Diffs you can review and comment on, then follow the pull request through CI.
  • A Browser pane that previews your running app while Claude checks its own changes.
  • Panes for chat, diff, browser, terminal and a file editor arranged side by side.
  • Side questions that use the session's context without derailing the main task.
  • A choice to run on your machine, in the cloud or over SSH.

The + button next to the prompt box adds file attachments, skills, connectors and plugins, the same building blocks you'd use elsewhere in Claude.

Local, cloud or SSH

The environment selector is where you choose where the work happens. Local runs on your computer, in the folder you open. SSH runs on another machine you connect to. Cloud runs the session on Anthropic's infrastructure, and Anthropic's quickstart says cloud sessions continue even if you close the app. A cloud session asks you to pick repositories instead of a folder.

One catch: the modes available differ by environment. Cloud sessions support Accept edits, Plan and Auto, and Bypass permissions isn't available there at all. Pick Cloud only for work that already lives in a GitHub repository, because that's what a cloud session clones.

Desktop vs the command line

Most of Claude Code works the same in both, with a few differences Anthropic documents. The dontAsk permission mode is only available in the command-line version. Auto mode needs a recent model: Claude Opus 4.6 or later, Sonnet 4.6 or later, Haiku 5.5 or a Fable model. And organisation administrators can switch auto mode off for everyone with a managed setting called disableAutoMode, so if you can't see Auto at work, that's usually why.

In a cloud session, one small quirk: the @ shortcut for listing files only works after you've sent your first message.

Who it suits

If you already live in a terminal, the command-line version works fine and Anthropic's docs explain what carries over. The desktop app suits everyone else: founders building a small internal tool, marketers fixing a website template, anyone who wants to see every change before it lands. Being able to read a diff and click accept or reject is what makes it feel safe.

Work with me

Want AI doing the heavy lifting in your marketing?

I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.

A sensible first job is something you can check by eye. Ask it to tidy a messy folder of files and rename them consistently, or to turn a spreadsheet export into a clean one-page report. You'll see exactly what changed before you keep it.

Prefer to work in the cloud? Read Claude Code on the web.

My verdict

Claude Code desktop is the easiest way into Claude Code for anyone who isn't a developer. Keep it in Manual or Plan mode until you trust what it does, work on copies, and only move to Auto once you've watched it handle a few jobs well. For costs by plan, see Claude Code pricing.

If you'd like help choosing which AI tools your team should use and setting them up around the way you work, here's how we can work together.

Frequently asked questions

Is Claude Code available in the desktop app?

Yes. Claude Code is the Code tab at the top of the Claude desktop app's sidebar on Mac, Windows and Linux (beta). It requires a Pro, Max, Team or Enterprise subscription.

Do I need to install Node.js to use Claude Code desktop?

No. Anthropic says the desktop app includes Claude Code, so you don't need Node.js or the command-line version to use the Code tab.

What are the permission modes in Claude Code desktop?

There are five: Manual, Accept edits, Plan, Auto and Bypass permissions. They control whether Claude asks before editing files or running commands, and you can switch at any time.

Can Claude Code desktop run sessions in the cloud?

Yes. In the environment selector you can choose Local, SSH or Cloud. Cloud sessions run on Anthropic's infrastructure and continue even if you close the app.

Is Claude Code desktop on Linux?

Yes, in beta. Anthropic's docs say you install the Claude desktop app on Linux with apt.

Published and maintained by the Lilach Bullock team, covering marketing, AI and business growth.
Your buyers are asking AI who to use. Does it say you?

See for free whether ChatGPT, Claude, Perplexity, Gemini and Google name you, and get the plan to become the answer.

Check my AI visibility →
Sundays only

Get the Sunday newsletter.

One email a week. AI experiments, marketing tactics, and the workflows Lilach is building right now in her own business.

Subscribe free

Let’s get your marketing running on AI.

Book a free 30-minute call

We figure out what you need, where AI fits in, and what working together would look like.

Book the call →

Or take the 30-second calculator

You’ll see the hours and the money quietly leaking out of your week, and the three workflows worth building first.

Take the calculator →

Or grab the free AI resource library

Prompt packs, templates, checklists, and swipe files. The exact tools I build for paying clients. Yours, free.

Get the library →
Keep reading

More from the blog.