Claude Code Singapore: Step-by-Step Setup Guide for Non-Technical Professionals

This is the guide I wish existed when I first tried to set up Claude Code. There are plenty of tutorials written by developers for developers — full of assumptions about what you already know, peppered with jargon that gets no explanation, and built on the premise that you're already comfortable typing commands into a black screen. This guide is not that.

This one is written specifically for Singapore professionals who know exactly what they want to build — a client tool, an automation script, an internal dashboard — but have never touched a terminal window in their lives. We're going to go through every step, explain what each thing is in plain English, and flag the moments where most non-technical users get stuck so you can get through them instead of giving up.

If you're still deciding whether Claude Code is right for you at all, take a step back and read our post on what Claude Code is and why non-techies are using it first. If you're ready to install and start building, this is the right place to be.

What You Need Before You Start

Before you touch anything, make sure you have the following in place. This is not a long list, but each item matters.

• A Mac or Windows PC. Claude Code runs on both. The instructions differ slightly between them — we'll flag where they diverge.
• A credit card for an Anthropic API key. This is how Claude Code connects to Anthropic's AI. You're not signing up for a subscription — you're paying for usage as you go. In your first week of experimentation, you'll likely spend somewhere between $2 and $5 SGD in API credits. It's metered, like a prepaid top-up.
• Node.js installed on your computer. Node.js is a free software package that your computer needs in order to run Claude Code. Think of it like installing a font that a document needs before it can display properly — it's infrastructure, not something you interact with directly. We'll cover how to install it below.
• About 30 to 45 minutes of focused, uninterrupted time. The setup itself isn't technically hard, but it requires your full attention. Don't try to do this between meetings or while multitasking. Sit down, close your other tabs, and give it one complete session.

That's genuinely it. You don't need to know how to code. You don't need any prior technical background. What you need is the above list and the willingness to follow instructions carefully.

Installing Claude Code — Step by Step

Follow these steps in order. Do not skip ahead. Each step builds on the previous one.

1. Open Terminal (Mac) or Command Prompt (Windows). On a Mac, press Command + Space to open Spotlight, type "Terminal", and press Enter. On Windows, press the Windows key, type "Command Prompt", and press Enter. You'll see a window with a blinking cursor. It looks intimidating. It's just a text interface. Nothing will break unless you tell it to do something that breaks things — and we're not going to do that.
2. Install Node.js if you haven't already. Go to nodejs.org in your browser. You'll see a big green button that says "LTS" — that stands for Long Term Support, which means it's the stable version. Click it to download the installer. Once it downloads, open the file and follow the installation steps just like you would for any other software on your computer. Click through the prompts, accept the defaults, and let it install. When it's done, go back to your Terminal window.
3. Install Claude Code. In your Terminal window, type the following command exactly as written and press Enter: npm install -g @anthropic-ai/claude-code. You'll see a lot of text appear and scroll past — this is normal. It means Node.js is downloading and installing Claude Code. Wait for it to finish. You'll know it's done when you see your cursor reappear on a new line. This typically takes 1 to 3 minutes depending on your internet speed.
4. Get your Anthropic API key. Open your browser and go to console.anthropic.com. Create an account if you don't have one, or log in. Once you're inside the console, look for "API Keys" in the left-hand menu. Click it, then click "Create Key". Give it a name (something like "Claude Code personal" is fine), then click Create. A long string of characters will appear — this is your API key. Copy it immediately and paste it somewhere safe like a notes app, because you won't be able to see it again after you close this screen.
5. Connect your API key to Claude Code. Go back to your Terminal window. Type claude and press Enter. Claude Code will start up and ask you to enter your API key. Paste in the key you just copied. Press Enter. It may ask a couple of simple setup questions — answer them and press Enter for each one.
6. You're in. Once Claude Code is running, you'll see a prompt waiting for your first instruction. Type something and press Enter — Claude Code will respond. You are now running Claude Code on your computer. The setup is complete.

Your First Session — What to Type

The best way to understand Claude Code is to use it immediately. Here's a concrete first project that will show you what the tool can do without any risk of breaking anything important.

First, create a new folder on your Desktop. Call it my-first-project. On a Mac you can do this by right-clicking your Desktop and selecting "New Folder". On Windows, right-click and select "New Folder".

Now, back in your Terminal window, you need to navigate to that folder. Type the following command and press Enter:

cd ~/Desktop/my-first-project

The cd command stands for "change directory" — it's how you tell Terminal which folder to work inside. Now type claude and press Enter to start a Claude Code session inside your project folder.

When Claude Code is running, type this instruction:

Create a simple webpage that lists my top 5 clients and their contact details. Make it look professional, with a clean layout, a header with my business name, and each client in a card format.

Press Enter and watch what happens. Claude Code will ask you a few questions — your business name, the client names and contact details you want to include. Answer them in plain English. It will then create an HTML file and a CSS styling file inside your project folder.

When it's done, open your my-first-project folder on your Desktop, find the file called index.html, and double-click it. It will open in your browser as a working webpage. That's your first build. You didn't write a single line of code.

The Things That Will Confuse You the First Time

No matter how carefully you follow these steps, you will almost certainly encounter at least one moment where something looks wrong and you're not sure what to do. Here's a guide to the most common points of confusion — and how to handle each one.

The terminal looks broken when it's working. Lots of scrolling text, strange characters, or a long pause with no output — this is usually Claude Code doing its job. It's installing things, reading files, or running code. Unless you see an explicit error message with the word "Error" in it, assume it's working and wait.

Permission errors. If you see a message containing the word "permission" or "EACCES", your computer is blocking the action because it needs administrator approval. The fix on Mac is to type sudo before whatever command you were trying to run. For example, if npm install -g @anthropic-ai/claude-code gives a permission error, try sudo npm install -g @anthropic-ai/claude-code instead. It will ask for your Mac login password — type it and press Enter. Note: when you type a password in Terminal, nothing appears on screen. That's intentional. Just type it and press Enter.

Claude Code stops mid-task. It's probably waiting for your approval before taking an action. Look at the Terminal window — it may be showing a question or a prompt asking you to confirm something. Press Enter to approve, or type "yes" and press Enter. Claude Code is designed to pause and check before doing anything that modifies files, to make sure you're comfortable with what it's about to do.

Error messages appear in red text. Don't panic. Copy the entire error message, then type to Claude Code: "I got this error — can you fix it?" and paste the message in. Claude Code will read the error, diagnose the cause, and fix it. Error messages are data, not disasters. This is one of the genuinely powerful things about Claude Code — it can debug its own output.

What to Build in Your First Week

The biggest mistake new Claude Code users make is trying to build their full product on day one. The tool is capable of that, but you're not ready for it yet — not because of a lack of skill, but because you haven't developed the instincts for how to direct it well. Your first week should be about building those instincts.

Start with small, self-contained projects where you can see the result immediately and judge whether it's right. Good first-week builds include:

• A one-page website about your business or services, with a simple contact section
• A script that renames files in a folder according to a naming pattern you define
• A simple data table that reads from a CSV file and displays it as an HTML page
• A template generator that takes a few inputs and produces a formatted document

None of these are production-ready tools — they're exercises. The goal is to go through the cycle of describing what you want, seeing what Claude Code produces, giving feedback, and iterating. By the end of the week, you'll understand intuitively how to write instructions that get the results you want.

The principles that make Claude Code instructions effective are the same ones that make any AI instruction effective — be specific, describe the output not just the task, and include context about who will use what you're building. Our post on outcome-first prompting goes deep on this and is worth reading alongside your first week of Claude Code sessions.

When You're Ready to Go Deeper

Once you've completed your first few builds and you're comfortable with the basic workflow, there are several directions you can take from here.

For a real story of what it looks like to take a Claude Code build all the way to a live, deployed product — including the moments where things went wrong and how they were fixed — read our post on shipping to prod in 5 hours with Claude Code. It's the most honest account of the full build process we have, written by someone who had never shipped software before.

For the broader context of how Claude Code fits into the professional AI landscape in Singapore, and how to think about which tools belong in your workflow, see our post on what Claude Code is and why non-techies in Singapore are using it.

For those who want to go further into advanced Claude usage — understanding the architecture of how Claude thinks, how to structure complex multi-step projects, and how to get consistently high-quality output — the Claude Architect Exam Decoded post covers the deeper principles that separate casual users from power users.

And if you find yourself wanting to discuss any of this with other Singapore professionals at the same stage — people who are figuring this out alongside you, sharing what works and what doesn't — that's exactly what the Cowork SG community is for.