You vibe coded an app. Here is what is still missing before it is a business

A vibe coded app from Lovable or Base44 arrives further along than most first products ever got: it runs, it has login and a database, it is hosted, it can sit on your own domain, and both tools can wire a real checkout into it. What neither tool hands you is the business: what exactly you sell and for what price, the legal and tax reality behind that checkout button, the support when real users hit real problems, and above all the customers. This page is the honest map of that remaining distance, and what to do about it this week.

What do the vibe coding tools actually cover?

More than skeptics think, and it keeps growing. Lovable builds full-stack apps and now sets up payments natively: by its own documentation it creates the provider account, creates the products and prices, builds the checkout flow and registers the webhook endpoints, all from a prompt. Base44 generates the frontend and backend together, with an authentication system and a database in every published app, and hosting built in so an app is live and shareable as soon as it is ready. The building, which stopped a decade of would-be founders at "I need a technical cofounder", has genuinely collapsed into days. We say that without jealousy: our own product exists because that collapse is real, and we tested it on ourselves in public: we built a real business in 7 days with baget

So take the building off your worry list. The list that remains is the subject of this page.

What does the checkout button not do?

A checkout renders a price. It does not answer for it. Behind any real payment flow, whoever built it, sits verification of who you are, in some setups a registered entity once real money moves, tax handling that varies by country and by provider arrangement, payouts to an actual bank account, and a refund policy someone will eventually invoke. Some providers absorb a real share of this, acting as the seller of record and handling tax and invoicing; that is a genuine simplification, not a myth. What no arrangement absorbs: it is still you deciding what is sold, at what price, under what terms, with what promise. Checking what applies to your country and your setup is an afternoon of reading, and it belongs before the first real charge, not after. The startup math of this whole stage is small and we itemized it: do you need money to start a business?

Who decides what you are selling?

You, and the tools are honest about this: they build what you describe. But a side project's description usually answers "what can it do" and a business needs "what outcome does who pay for". Those are different sentences. When we ran our own seven-day experiment, the agents produced a working storefront and priced the offer; then the humans renamed it, repositioned it and repriced it, and those decisions took longer than the building and could not have come from the tool. Yours will be the same. One offer, one price, described in the buyer's words: that is job 2 finished, and most vibe coded apps have not finished it, because nobody made them.

Where do customers come from?

Not from the builder, and not from the launch. No feature of any tool, ours included, replaces finding out what people pay for and telling people it exists, again and again. The path is the same one every business walks: name who it is for, reach ten of them you can already reach, put a real price in front of them, and let the noes teach you. Weeks to months, and the offer changes at least once. The step-by-step of that stage, and of everything else this page compresses, is in the companion guide: how to turn a side project into a real business

What breaks first when real users arrive?

The operational week, because nobody vibe codes that. A stranger's data hits an edge case your own use never found. Someone wants a refund on the day your job runs late. Two support emails become nine. None of this is a catastrophe; all of it is the actual texture of a running business, and it is why "published" and "running" are different states. Decide before you invite strangers: where support lands, how fast you honestly reply given the rest of your life, what you do when the app is wrong and money was taken. Written down once, these fit in a page. Improvised, they eat the exact hours that were supposed to grow the thing.

What should you do this week?

In order. Write the one sentence: who pays, for what outcome, at what price. Do the legal and tax afternoon for your country before the first real charge. Put the offer in front of ten reachable people as a price, not a survey. Decide where support lands. And keep building only what the asks demand, which will be less than you want to build. That sequence, not more features, is what moves a project across the line.

The app exists. Describe the offer in one sentence and see the business side stood up around it. It costs nothing to look.