You vibe-coded an app. Now what?

You got the app built. Claude Code or Cursor or Lovable did the part that used to require a technical cofounder, and it works — you can click through it, it does the thing, maybe a few friends have poked at it.

And then the momentum stops, because the next task isn't a task. It's a fog of twelve half-known obligations: incorporate? terms of service? pricing? a landing page? talk to users? Some of those are urgent, most aren't, and nothing on the internet tells you which is which, because the people writing about startups are mostly writing for people who already know.

Here's the honest ordering.

The only three things that are actually urgent

1. Find out whether anyone wants it. Not whether people say it's a cool idea — whether anyone will give you something that costs them: money, a work email they check, or thirty minutes on a call. Everything else on this page is wasted effort if the answer is no, and you can find out in a week.

The trap here is that you already built it, so validation feels backwards. It isn't. Building it made you feel validated; it didn't produce evidence.

2. Don't create a legal mess you can't undo later. There is a short list of things that are cheap now and expensive-to-impossible later:

• A contractor writing code with no IP assignment. Their code is theirs. This is the single most common unfixable problem in early startups, and it surfaces during diligence, years later, when you can no longer find them.
• A cofounder with no vesting. If they leave in month three, they keep their half forever.
• Collecting user data with a copy-pasted privacy policy that describes a different product than yours.

None of these require incorporating tomorrow. They require not doing the specific wrong thing.

3. Be able to say what it is in one sentence. Not the mechanism — the change. "AI-powered workflow automation for teams" describes what's in the box. "Your standups write themselves and the summary lands in Slack by 10am" describes what's different on Monday.

If you can't do this, everything downstream — the landing page, the cold email, the Reddit comment, the pitch — is harder than it needs to be, and you'll blame the channel.

The things that feel urgent and aren't

Incorporating. Real trigger, not vibes: you're taking money, you're signing something, you have a cofounder, or an investor is asking. Absent those, an entity is annual paperwork protecting nothing yet. See When should you actually incorporate?.

The logo, the name, the brand. A placeholder name costs nothing until people start referring you to each other by it — which, if you're honest about where you are, hasn't started.

Choosing a pricing tier structure. You need a price, because a price tells you who your customer is. You do not need three tiers with feature gates before you have ten customers.

Scaling anything. Your database is fine. Your hosting is fine. The bottleneck is that nobody knows you exist.

The things nobody warns you about

Your app being "done" is not a milestone anyone else can see. You experienced weeks of progress; the world experienced nothing. The first month after launch feels like failure specifically because the feedback loop that was tight while building — write code, see it work — has been replaced by a much slower one.

AI will answer any question you ask it, which is the problem. Ask a chatbot "should I incorporate?" and you'll get a competent general answer. What you won't get is the question you didn't know to ask — the one about where your customers are located, or about the contractor, or about what happens to the entity if you move countries. Answers are cheap now. Knowing what to ask is the part that's still scarce.

The app you built still needs keeping alive, and you can't read the code. That's a solvable problem with its own set of habits — memory files for your AI, standing rules, monthly check-ups. It's a different topic from this page, so it has its own series.

Most of these decisions are reversible, and a few are not. Almost all founder anxiety is spread evenly across both categories. Sorting them is most of the value of thinking it through at all.

A sequence that works

1. Talk to five people who have the problem and are not your friends. Write down the words they use.
2. Write the one sentence. Test it on someone who's never heard of your product.
3. Ship a way to pay, even a bad one. A Stripe link is enough. The gap between "would use it" and "did pay" is the entire game.
4. Fix the unfixable things — the contractor agreement, the vesting, the data policy — as they become real. Not before, not after.
5. Then incorporate, pick the name, build the tier structure, and worry about scale.

Working it through properly

Each of the steps above is a Groundwork project type — a structured path through the questions in order, with an AI that builds context about your specific situation as you go, and ends in something real rather than advice.

Reasonable places to start, depending on which part of the fog you're standing in:

• Product Validation — if you don't yet know whether anyone wants it
• MVP Scope — if it's built but sprawling
• Incorporate Now — and As What? — if the entity question is the one nagging
• Go-to-Market Strategy — if it works and nobody's using it

Or browse all of them.