How to Use Claude Code If You're Not a Developer

How to use Claude Code if you're not a developer, quick answers (as of November 4, 2026):

You don't have to be a developer.
You need a paid Claude plan (or a pay-as-you-go Anthropic API account). Pro, at $20 a month ($17 a month billed yearly), is the cheapest plan that includes Claude Code. The free plan doesn't include it.
My four steps: install it in the terminal, set up a folder structure, give it a database, and add the Claude in Chrome extension.

So yes, I'd call myself a non-developer. I have a little bit of experience, but not much. I've never had a job as a developer, never freelanced as a developer, nothing like that.

I actually started with Lovable around the beginning of the year. I had some developers who were supposed to be building an app for me, and I was trying to get them to use Lovable, thinking it would help them go faster. They started out of the gate really, really well. Within a couple of days, it was clear they weren't going to make very much progress. So I just jumped in and tried to do it myself, and found out I could actually do a ton with it, including connecting to APIs and all kinds of stuff.

I did that for a month or two. Then I realized Claude Code directly would probably be cheaper, so I switched to Claude Code. I started using it to try to build some features and ship them to my development team to review. Eventually they became the bottleneck. So I ended up just shipping stuff myself.

If you're just getting started with Claude, I'd say do what I'm doing. Here's how to use Claude Code if you're not a developer, in the order I'd do it.

Step 1: Install Claude Code in the terminal

The first thing you're going to need to do is install Claude in the terminal. I use PowerShell. Claude can walk you through how to install it, but here are the commands, straight from Anthropic's setup page:

Windows (PowerShell): irm https://claude.ai/install.ps1 | iex
Mac or Linux (Terminal): curl -fsSL https://claude.ai/install.sh | bash

Then open the terminal, type claude, and here we go.

On Windows, it needs Windows 10 version 1809 or newer, and you don't have to run it as administrator. Anthropic also recommends installing Git for Windows, but it's optional. Without it, Claude Code runs its commands in PowerShell.

First, understand that there are some risks with AI. It can make mistakes. There's always that.

The reason I run it on my desktop computer is that it can look at literally everything on it, and that's fine for me. If you don't want that, proceed at your own pace and with your own risk.

For me, it helped me troubleshoot why my computer was glitching. My computer used to glitch every 3 days or so. I'd have to restart, I'd lose all my work, and I'd have to reopen everything. Claude Code helped me troubleshoot that, update the drivers, test things and get it all working. I don't think it's crashed since.

Claude Code vs Claude Desktop: the terminal or the app

Using it in the terminal rather than inside the app has benefits, because it can go outside of the repo more easily. You can also get started without a repo. You can start with just a folder, and then use Claude Code to help you set up the repo.

When I was using Cowork, I could select a folder, and that was fine. But it would ask for permission a lot for different stuff, and it seemed easier to get it not to in the terminal.

Fact check (November 4, 2026): I thought the Claude desktop app made you select a repo first. It doesn't. The Code tab in the desktop app asks you to open a project folder. You only need Git if you want each session to run in its own separate copy of the project. Older Windows versions of the app did ask for Git before any session, so update the app if you see that.

Step 2: Set up a folder structure

Second, I'd have it set up some kind of folder structure, so everything doesn't just get put everywhere all willy-nilly.

This is the first fork in the road, depending on whether this is for a company or for somebody using it for a little bit of stuff here and there. The more robust the use, the more important it becomes to have your data structured.

By default, Claude likes to write a bunch of MD (markdown) files. Even if it keeps them organized in the folder structure you set up, they're still files that have to be read. And you'll find out quickly that Claude does not always follow instructions. Anything that's written down is open for interpretation and skimming, because it's trying to save you tokens on context. At least that's my understanding.

Fact check: Anthropic's own docs say something close. Claude treats its instruction files "as context, not enforced configuration." Long instruction files make it follow them less reliably, and most other files only get read when Claude decides it needs them.

Step 3: Give Claude Code a database

The next thing would be to connect Claude Code to a Supabase database, so you have somewhere you can put data for free, and your data is in a better structure.

If you don't do this, Claude will use whatever local tools are available, to some extent, like if you upload a spreadsheet or try to get it to process one. But in my opinion, it's been enormously helpful to give it a way to structure the data.

For light use, that could just be connecting it to Google Sheets, so it can create Google Sheets. But if you're dealing with lots of data, you're going to blow up a Google Sheet. So you're really going to need a database. Supabase is free to start with, so that's what I'd recommend.

Supabase's free plan (as of November 4, 2026): $0 a month, a 500 MB database and up to 2 active projects. Free projects pause after a week of inactivity. Paid plans start at $25 a month. Supabase pricing

Step 4: Install the Claude in Chrome extension

The last thing I'd do is install the Claude in Chrome browser extension. That way, whenever you need to do something Claude can't web fetch, it can just open your browser and do that thing, for the most part.

It works in Chrome and Edge, and Claude Code also connects to it in other Chromium browsers like Brave. It needs a paid Claude plan. Start Claude Code with claude --chrome, or type /chrome in a session to connect it.

Claude refuses to log into stuff. It will open your browser and drive the window after you log in, to do pretty much anything. But it won't log in for you. So it's always blocked on you for logging in, even if you give it the password or give it access to your passwords. Anthropic's docs say the same: when Claude hits a login page or a CAPTCHA, it pauses and asks you to handle it.

That's where Grok Bot comes in for me. One of the biggest reductions in friction in using Grok Bot has been using it for logging into stuff.

How to use Claude Code: FAQ

Can you use Claude Code if you're not a developer?

Yes. I'm not a developer, and I ended up just shipping stuff myself with Claude Code.

Is Claude Code free?

No. Pro is $20 a month ($17 a month billed yearly) and is the cheapest plan that includes Claude Code. The free plan doesn't include it. Prices as of November 4, 2026.

Do I need to use the terminal?

No. The Claude desktop app has a Code tab that runs Claude Code without a terminal. I use the terminal because it can go outside of the repo more easily, and you can start with just a folder.

Do I need a Git repo to use Claude Code?

No. You can get started with just a folder, and then use Claude Code to help you set up the repo. The desktop app asks for a project folder, not a repo.

Does Claude Code work on Windows?

Yes. It installs with one line in PowerShell on Windows 10 version 1809 or newer.

Can Claude Code log into websites for me?

No. With the Claude in Chrome extension, it can drive your browser after you log in, but it stops at login pages and CAPTCHAs and asks you to handle them.

More from me

I'll come back and update this when or if I have more to say while using Claude Code.

If you want to follow along, follow me on X, or get my newsletter, Running with Robots.

-Matt