Claude Code first run. Login, folder trust and your first day

What happens the first time you start Claude Code. Which accounts work, how login goes, what the folder trust prompt means and how to spend your first day with it.

Part of the guideClaude Code, the complete guide. From installation to your own skills and plugins

In this article
  1. Which accounts work with Claude Code
  2. First start and login
  3. The folder trust prompt
  4. Your first day with Claude Code

In short

On the first run, Claude Code asks you to log in through the browser, and in a new folder it asks whether you trust it. We cover which accounts work, where login errors come from, what folder trust changes and how to get through the first day calmly.

Someone has an active Claude subscription, logs in without trouble, and the very first request fails with an error saying the organization has been disabled. The cause sits in their shell config. An old API key from a previous project takes priority over the subscription. It is one of several things worth knowing before the Claude Code first run turns into a debugging session. Everything below was checked against the Anthropic documentation as of 10 November 2026.

The first ten minutes with /init, /context and /usage are already covered in What is Claude Code. Here we go one step earlier and one step further. The whole series lives in the Claude Code guide.

Which accounts work with Claude Code

You can log in with several kinds of account.

AccountHow you log in
Claude Pro, Max, UltraYour claude.ai account in the browser
Claude Team, EnterpriseThe claude.ai account your admin invited
Claude ConsoleConsole credentials, after an admin invites you
Amazon Bedrock, Google Cloud, Microsoft FoundryEnvironment variables, or the wizard under 3rd-party platform

The free claude.ai plan does not include Claude Code. The first time you log in with a Console account, a workspace called Claude Code is created in the Console, where you can track costs. A company cloud account does not need a browser login.

First start and login

Go to your project folder in the terminal and start Claude Code.

bash
cd ~/projects/shop
claude

On the first start, Claude Code opens your browser. You sign in, and the terminal shows Login successful and asks you to press Enter. If the browser does not open, press c to copy the login address. If the browser shows a code instead of sending you back, paste it at Paste code here if prompted. Logging in over SSH, in WSL or in a container works the same way.

Your login is saved, so later starts do not ask again. On a Mac it goes to the Keychain, on Linux and Windows to a .credentials.json file in the .claude folder in your home directory. Three days before the login expires, Claude Code warns you at startup, and /login renews it.

Now the story from the top. If ANTHROPIC_API_KEY is set in your environment, Claude Code asks whether to use it. Once you say yes, the key wins over your subscription. An old key from a disabled organization then causes errors even though the subscription is fine. Run unset ANTHROPIC_API_KEY, remove the export ANTHROPIC_API_KEY line from ~/.zshrc or ~/.bashrc and start again. /status inside a session shows which login is active.

/login switches accounts, and /logout signs you out and walks you through the whole setup again on the next start. To keep a personal and a work account apart, give each its own config folder through the CLAUDE_CONFIG_DIR variable, for example in a claude-work alias.

The folder trust prompt

In a folder you have not trusted yet, Claude Code shows a trust dialog. It is not a formality. A project can ship its own .claude/settings.json with rules that let Claude act without asking, plus extra directories. Those rules only start working after you accept, and the dialog lists them so you can read them first. Deny and ask rules apply right away, because they only restrict.

Servers in a project's .mcp.json have their own approval prompt. In someone else's repository, read both before you click on.

One exception surprises people. If you start Claude Code directly in your home directory, trust holds for that session only and the prompt returns on every launch. There is no setting to change that. Start Claude Code in a project folder, where trust is saved for good.

Two more things can appear on the first start. Claude Code offers to set up your terminal, and in Apple Terminal on a Mac that turns on the Option key as Meta. It also shows a one-time notice that the session starts in auto mode. What that means is covered in Claude Code permissions.

Your first day with Claude Code

Do not hand it a big job on day one. It is better to see how it reads your project and what working with it feels like.

  1. Ask about the project, for example "what does this project do" or "where is the entry point". Claude reads the files it needs, you do not paste anything.
  2. Type /init. Claude goes through the project and writes a first version of CLAUDE.md. With CLAUDE_CODE_NEW_INIT=1 set, it also walks you through skills, hooks and personal memory files. How to trim that file is in our article on CLAUDE.md.
  3. Before the first change, press Shift+Tab until the status bar shows manual mode on. In this mode Claude asks before every file edit and command, and you choose Yes or No.
  4. Give it a small task with a clear finish, for example "add validation for an empty email field in the signup form". Specific beats "fix the form".
  5. Look at the changes with /diff. It shows your working tree changes against git, including Claude's edits, and lets you view each turn on its own.
  6. If the change is good, ask for a commit with a message. If not, say what to fix, or roll it back with /rewind.

A task like this takes a few minutes and shows the whole loop. A question, reading files, a change, a review and a commit. If something goes wrong with the install or login, claude doctor in a plain terminal checks the install and settings without starting a session.

If Claude Code is not installed yet, start with the guide for macOS, Windows or Linux. The next step is the commands and shortcuts you will use most.

Spend the first day in Manual mode on small tasks, and read the trust prompt in every new repository. If you want to see where AI could take work off your business, our free AI audit is a good place to start.

What would you like to change in your business?

Leave your number. We will talk about your process and see whether we can help. The first call is free.

When is the best time to call?
Controller: Oleksandr Bondarenko (BondSoft)Privacy PolicyTerms of Service
Legal details

The data controller is Oleksandr Bondarenko, trading as Oleksandr Bondarenko Software Solutions (BondSoft), ul. Jana Heweliusza 11, 80-890 Gdańsk, Poland, NIP 5833576264, e-mail: office@bondsoft.pl.

We process your data solely to contact you about your enquiry and to present an offer — on the basis of your consent (Art. 6(1)(a) GDPR and Art. 398 of the Polish Electronic Communications Law) and our legitimate interest in handling the enquiry (Art. 6(1)(f) GDPR).

Providing the data is voluntary but necessary for us to reply. You may withdraw your consent at any time by writing to office@bondsoft.pl — this does not affect the lawfulness of processing before the withdrawal.

You have the right to access your data, rectify it, erase it, restrict its processing, transfer it, and lodge a complaint with the President of the Polish Personal Data Protection Office.

Full information: Privacy Policy.

Send us an email

The service this text is about

Automation and AI / OCR

Automation of document re-typing, OCR and checking rules, and your data stays with you.