Vibe coding for beginners.
You do not need to learn to code first. To start vibe coding, pick one small annoyance in your week, open one AI tool, describe what you want in ordinary English, and change one thing at a time until it works. Most beginners have something live on the internet within a weekend.
Yes, you can do this. Here is why.
Almost everyone arrives here quietly convinced they are the exception. That everyone else has some background they are missing. They do not.
Until about 2025, building software genuinely did require years of study. That was true, and it kept a lot of people out. It stopped being true recently enough that most people have not updated their sense of what is possible.
What the job actually is now: describe what you want, look at what came back, and say what is wrong with it. You already do this. You do it with contractors, with designers, with anyone you have ever asked for something. The only new part is that the thing you are describing gets built in ninety seconds.
People walk in curious and leave with something they built.
The people who do well here are not the technical ones. They are the ones who pick something small and finish it. Finishing a tiny thing teaches you more than starting an ambitious one.
Four things, and you probably have three.
Any laptop from the last several years. The AI runs on someone else's computers, so yours only has to display a web page.
Claude, Lovable, Replit, or Cursor. Pick one by reading nothing and choosing whichever name you like. Seriously. Any of them work, and choosing is the part where people stall for three weeks.
Not three hours of study. Three hours of poking at something. You can split it across evenings if a weekend is not realistic.
Something in your own week that is mildly irritating and could be fixed by a simple tool. This is your project. Not the big idea. The annoyance.
Notice what is not on this list: a course, a degree, a technical friend, a co-founder, or permission. If you want the full picture of what this practice is before you start, read what is vibe coding.
Friday night to Sunday night, hour by hour.
This is the plan we give people who have never built anything. It is deliberately unambitious, because the goal of your first build is not the app. It is proving to yourself that the loop works.
Friday evening: pick the annoyance
45 minDo not pick your big idea. Pick something small that irritates you every week. Write one sentence about what it does. Then close the laptop.
Saturday morning: get something on screen
2 hoursOpen one AI tool. Describe the simplest possible version and let it build. Do not judge it yet. The goal this morning is only to see something that exists and responds when you click it.
I have never built software before, so explain things simply as we go. I want to build the simplest possible version of this: [one sentence about your thing] One screen. No accounts, no payments, nothing fancy. Build it, then tell me in plain English what you made and how to try it.
Saturday afternoon: make it yours
2 hoursChange one thing at a time. Wording, colours, the order of the fields, what happens when someone submits. After each change, use it. This is the afternoon where it starts feeling like a real thing.
Sunday: put it on the internet
1 hourAsk your tool to deploy it and give you a live link. Open that link on your phone. This is the moment most people describe as the one where it clicked.
Deploy this so it has a real web address I can open on my phone and send to a friend. Walk me through anything you need me to do, step by step, assuming I have never done this.
Sunday evening: show one person
10 minSend the link to somebody. A friend, your group chat, the community. Not for feedback. For the fact that you made a thing that exists at an address, which was not true on Friday.
Five first projects that actually finish.
Every one of these is small enough to complete in a weekend and useful enough that you will keep it. Steal one.
Your work, your side business, your services. One page, a description, a way to contact you. Boring and genuinely useful, and you will finish it.
Bookings, enquiries, orders, sign-ups. If you currently handle any of these through DMs, this project pays for itself the first week.
Workouts, reading, expenses, plants you keep forgetting to water. You get a real database and real saving without it feeling like a lesson.
Quotes, dosages, splitting costs, unit conversions nobody else needs. If you keep a formula in a spreadsheet, it wants to be a small tool.
Recommendations, a team wishlist, a club's events. This one teaches you how several people share the same thing, which is a genuinely useful step up.
If none of those fit, we keep a free pack of 100 mini app ideas with the prompts to build them, which exists for exactly this moment.
Six things that stop beginners.
Every one of these is something we watch happen weekly. Knowing them in advance genuinely saves you a month.
Your first build should be disposable. Starting with the idea you have carried for years means every rough edge feels like failure, and you quit. Build something you do not care about first.
There is always another tool. Reading about eleven of them is not progress, it is a comfortable way to avoid starting. Pick one, badly, today.
Ten features in one request gives you ten things to debug at once. One change per message, every time. This is the habit that separates people who finish from people who do not.
AI is cheerfully confident about work it did not do properly. Click the button yourself. Every single time. It takes ten seconds.
Two weeks on the colour scheme of a thing that is not on the internet yet. Deploy it rough. You can keep improving it at a live link, and the live link changes how you feel about it.
You will hit something that makes no sense and feels like proof you were not cut out for this. It is not. Everyone hits it, usually in the first weekend. Start a fresh chat, describe where you are, and keep going.
Nobody does the first one entirely alone.
Two things get beginners unstuck faster than anything else, and both are free.
Ask the AI to teach you, not just build
Most beginners forget that the thing building their app is also a patient tutor with no opinion about what you should already know. Try: “Explain what you just did, in plain English, like I have never seen code.” Do that a few times a session and you will pick up more than a course would give you.
Sit near other beginners
Building alone makes every wall feel personal. Watching five other people hit the same wall on the same afternoon reframes it as normal, because it is. Our community is free to join, and the weekly newsletter walks through a new build in plain English every week. If you would rather have it all in one place, the book Hey Claude is $14.99.