Vibe-coded apps that make money: what they did after launch

The pattern in one paragraph

AI-built products that make money tend to share four habits: they sell to a narrow, specific buyer; they charge from the start; the founder talks to customers directly; and they grow through one channel worked deeply before adding another. The tool used to build them barely matters.

Public examples

All figures are reported by the founders or by the publication named, as of the date shown. They are founder-reported, not audited, and some may have changed since.

What each one teaches

fly.pieter.com: charge on day one. Pieter Levels built a browser flight game with AI tools in hours. But the game itself was free, so the money came from two simple offers: a paid jet and ad space on in-game blimps. So revenue started within days, not months. The lesson is not “go viral”. Levels already had a large audience on X, which most founders don’t. The lesson is to put a price on something from the first day.

Launch Fast: borrow an audience. Hasaam Bhatti built a tool for Amazon sellers in about 48 hours with Cursor. But he did not launch to strangers. He went to a coaching program whose students were exactly his buyers and offered a partnership. So he had paying customers on day one, and his MRR grew from $10K at day 30 to $30K later. The lesson: find a group that already gathers your buyers, and give them a reason to introduce you.

Setter AI: sell before you build. The founders put up a landing page and a demo for a product that did not exist yet. Visitors booked calls, and some paid a refundable $500 deposit for early access. So they knew people would pay before they wrote most of the code. Later customers came from SEO, free tools, comparison articles and YouTube.

Habit Pixel: small price, slow and steady. A solo developer launched a habit tracker in May 2025 and posted about it on X and Reddit. But at $1.99 a month, each customer is small, so growth came from many small steps: building in public, a Black Friday deal, a launch on a product directory, app store work and a few dollars a day in ads. It took about eight months to pass $1K MRR. That is a realistic pace for a low-price consumer app.

Base44: share the build in public. Maor Shlomo built an AI app builder on his own and grew it mostly through word of mouth, sharing the process on LinkedIn and X. It reached 10,000 users in its first three weeks and was sold for $80M about six months later. Most founders will not have this outcome. But the habit behind it (showing the work where buyers already read) is free for anyone.

What they did after launch

They picked a buyer, not a market. “Freelance designers who send proposals” instead of “small businesses”. A narrow buyer makes every message sharper.

They charged early. A real price separates people who need the product from people who are curious. Monthly pricing first; annual plans later, once customers trust the product.

They sold in conversations. First customers came from replies, DMs, communities and calls, not from a launch day.

They worked one channel deeply. One community, one content format, or one outreach list, every week, until it worked.

They tracked where customers came from. So they knew which channel to double down on.

The pattern, side by side

None of these founders won because of the tool they built with. Each won because they knew who would pay, asked for money early, and kept showing up in one place where those buyers already were.

Two moves you can copy this month

Most of the examples above are hard to repeat. You can’t borrow Pieter Levels’ audience or plan for an $80M exit. But two of the moves need no audience at all. So they are the best place for a new founder to start.

Move 1: Borrow an audience (the Launch Fast move)

Launch Fast did not launch to strangers. It went to a coaching program whose students were exactly its buyers. You can do the same on a smaller scale.

1. List the groups that already gather your buyers. Paid communities, coaches, course creators, newsletters, podcasts, associations, software partners. Aim for 20 names.
2. Pick the five with the closest match. The best partner serves your buyer but does not sell what you sell.
3. Offer something useful to their members. A free month, a group discount, a live session, or a revenue share. Make it easy for the partner to say yes.
4. Send a short note. “Your members ask about [problem] a lot. I built [product] to fix it. I’d like to offer your members [offer]. Would a 15-minute call make sense?”
5. Track every lead from the partner. Give each partner its own link, so you know who sent paying customers.

One good partner can bring more first customers than a month of posting.

Move 2: Sell before you build (the Setter AI move)

Setter AI put up a landing page and a demo before the product existed, and some visitors paid a refundable deposit. That turns “would you use this?” into “will you pay for this?”.

1. Write a one-page offer. Who it is for, the outcome, the price, and when it will be ready.
2. Show a demo. A short video or a clickable prototype is enough.
3. Ask for a small, refundable deposit or a pre-order. Say plainly that it is refundable.
4. Talk to everyone who pays. These are your design partners. Build what they need first.
5. Set a deadline. If nobody pays in two to three weeks of real outreach, the offer or the buyer is wrong. Fix that before you write more code.

A deposit is the strongest early signal you can get. It costs the buyer something, so it means something.

Pricing your app

• Start with one or two simple plans.
• Name plans for who they are for (“Solo”, “Team”) rather than “Basic” and “Pro”.
• Charge monthly first; it is an easier yes for a new customer.
• Raise prices once you have customers who would be upset to lose the product.

A worked pricing example

Say you built a tool that helps freelance designers send proposals faster. You talked to ten designers. Six said proposals take them two hours each, and they send four a month.

That is eight hours a month. But if you price by the hour you save, you will overthink it. So start with a simple test: one plan at a price that feels slightly uncomfortable to say out loud, and watch who pays.

• Week 1: offer the plan to the six designers who told you about the pain.
• Week 2: if most say yes without pushing back, the price is too low. Raise it for the next ten.
• Week 3: if most say no, ask why. If the answer is “not now”, the price may be fine and the timing is wrong. If the answer is “too much”, test a smaller plan.

Setter AI took this further: a refundable $500 deposit before the product existed. That is the strongest signal you can get that a problem is worth solving.

Free plans and trials

A free plan brings users. But users who never pay can hide the fact that nobody wants to pay. So most founders at this stage are better off with a short free trial, or with no free tier at all, as Launch Fast chose. Add a free plan later, when you know it leads to paid accounts.

Common mistakes after launch

Most AI-built products that stall make the same few mistakes in the first month.

1. Building more instead of selling. New features feel like progress. But with no paying users, nobody can tell you which feature matters.
2. A buyer that is too broad. “Small businesses” is not a buyer. “Event rental companies that send quotes by hand” is.
3. Waiting for a launch day to work. A launch brings visitors for a day. First customers come from conversations the founder starts.
4. No price. Free users tell you the product is interesting. Paying users tell you it is useful.
5. No tracking. Without a source on each signup, you can’t tell which channel to keep. See the guide on UTM tracking for startups.
6. Losing the people who do show up. An inquiry with no reply for two days is a lost customer. Set up a simple follow-up rule before you add traffic.
7. Copying a viral story. Pieter Levels and Base44 are real, but they are outliers. Plan for the Habit Pixel pace and be pleased if it goes faster.

Can you sell a vibe-coded app?

Yes. Buyers of small software businesses care about revenue, customers and how much work the product needs, not about how the first version was written. But they will look closely at security, data handling and code quality. Check sign-in, payments and how customer data is stored before you charge anyone, and long before you try to sell.

What a buyer or a careful customer will check

Before you charge, and long before you try to sell the business, walk through this list:

• Sign-in and accounts. Can one user ever see another user’s data? Test it with two accounts.
• Payments. Are payments handled by a payment provider, with no card data stored on your side?
• Secrets. Are API keys kept out of the code and out of the browser?
• Backups. If the database is lost today, what do you lose?
• Revenue proof. Can you show revenue from your payment provider, not a screenshot?
• Customer concentration. Does one customer make up most of the revenue?
• Maintenance. Could someone else run it? Write down how it is deployed.

AI coding tools write working code fast. But working is not the same as safe. So spend one day on this list before your first paying customer, and you avoid the worst surprises later.

Your next 30 days

If your AI-built app has no users yet, start with the guide “You built it with Claude Code. Nobody is using it.”, which has a week-by-week plan.

A simple first-month plan

This is the same pattern the five examples followed, sized for a founder without an audience.

Checklist: the first 30 days after launch

Use this alongside the weekly plan above.

• I can name my buyer in one sentence.
• I have talked to at least ten people who fit that sentence.
• My product has a price, even a small one.
• Every signup or inquiry records where it came from.
• Every inquiry gets a reply within one business day.
• I picked one community or partner and show up there every week.
• I asked my first users what almost stopped them from buying.
• I wrote down which channel produced each paying customer.
• I stopped adding features nobody asked for.
• I checked sign-in, payments, secrets and backups.

How to tell it’s working

If two or more signals point to “change course”, go back to the buyer and the offer before you add traffic.

Questions and answers