Any vibe-coded apps still making a lot of money?

Why is it harder to tell whether vibe-coded apps are making real money now?

Yes, vibe-coded apps are still making serious money today, but we finally have enough history to separate short-lived launch spikes from businesses that kept growing.

The first wave of vibe coding produced spectacular screenshots. Pieter Levels' Fly reached roughly a $1 million annualized revenue pace almost immediately. That proved a product built extraordinarily fast with AI could attract paying customers. What happened afterward was more useful: Levels now places Fly in the "Okay" category on his public project history, which he defines as projects that took off and made money but were not sustainable.

Meanwhile, some newer products have lasted beyond their launch week. RevenueCat says Payout went from zero to roughly $80,000 MRR in four months. Launch Fast moved from $10,000 MRR after its first month to roughly $30,000 a few months later. Those trajectories tell us far more than a single Stripe screenshot.

The question has changed. We already know people will pay for software largely written by AI. We now want to know whether enough of those apps can keep customers, survive acquisition costs and turn fast building into a durable business.

Several have. That part is becoming much harder to argue with.

What should actually count as a vibe-coded app?

We count a vibe-coded app when AI generated a large share of the working product and prompting AI was central to the development process.

That definition needs to stay fairly strict. Almost every software company uses AI coding tools now, so counting every startup whose engineers use Copilot would make the category useless.

The clearest cases are founders who describe AI as their main implementation layer. Gil Hildebrand says Claude Code now writes roughly 90% of Subscribr's code. Stanley's founders explicitly describe building its original LinkedIn product in a 14-day vibe-coding sprint. Prayer Lock founder Mau Baron says he coded his initial mobile app with AI in three days despite coming from a finance background.

Experience does not disqualify someone either. Hildebrand had been programming professionally for roughly 25 years before Subscribr. Vibe coding describes how the product was implemented, rather than how little the founder understands software.

We also have to separate the apps from the companies selling the coding tools. Lovable recently reported more than $500 million in annualized revenue and now says roughly 60 million projects are hosted on the platform. That proves enormous demand for vibe coding itself. It tells us much less about how much money those 60 million projects make.

Lovable's own Build Economy survey is more useful on that point. Among respondents, 10.7% said they already made money directly from a product and another 9.1% earned from a mix of products and services. A much larger 60.5% planned to monetize but had not done so yet.

So we focus on the smaller group where customers are demonstrably paying for the resulting software.

Which vibe-coded apps are making serious money right now?

Yes, several named vibe-coded apps currently sit well above $10,000 a month, with the strongest public example at roughly $3 million ARR.

Stanley is the biggest clean example we found. Stan cofounder John Hu recently disclosed around $3 million ARR across Stanley's LinkedIn and Instagram products.

RevenueCat says Payout reached approximately $80,000 MRR in four months after its initial app was built through Claude Code and Cursor without manually written code.

Subscribr is also well into serious-business territory. Indie Hackers reported more than $62,000 a month after 18 months, while founder Gil Hildebrand described the business as being on track for roughly $1 million a year.

Lovable-built Lumoo reported €700,000 ARR within nine months. ToneAdapt founder Kyan disclosed approximately $25,000 of revenue across web and mobile during a recent four-week period. Prayer Lock reached roughly $21,000 in monthly sales within six months. Launch Fast recently reported $30,000 MRR.

None of those figures alone proves a permanent business. Together, they kill the idea that commercially successful vibe-coded apps are limited to $500 side projects.

Vibe-coded app	Recent public revenue	What was AI-built	Evidence quality
Stanley	~$3M ARR	Original product built in a 14-day vibe-coding sprint	Founder disclosure in recent interview
Payout	~$80K MRR	Initial design, code and assets produced through Claude Code and Cursor	RevenueCat documentation plus later RevenueCat update
Subscribr	>$62K/month reported	Founder says ~90% of code is written with Claude Code	Founder disclosure reported by Indie Hackers
Lumoo	€700K ARR	Product described as 100% powered by Lovable	Company case study with named customers
Launch Fast	~$30K MRR	Initial product built entirely through Cursor	Founder interview
ToneAdapt	~$25K over recent four weeks	Founder describes the product as vibe coded	Revenue dashboards shown in Starter Story interview
Prayer Lock	~$21K monthly sales at disclosed checkpoint	Initial app coded with AI in three days	Founder interview with revenue analytics

Is Stanley really a $3 million ARR vibe-coded product?

Yes, Stanley genuinely reached roughly $3 million ARR after starting as a 14-day vibe-coded build, although Stan gave the product an enormous commercial head start.

John Hu says the first Stanley product, an AI tool for LinkedIn content, reached roughly $200,000 ARR within six weeks. It crossed $1 million ARR several months later. Stan then launched an Instagram version, bringing the combined Stanley product line to approximately $3 million ARR.

That is a much stronger trajectory than one viral launch week. Revenue increased by roughly 15 times between the six-week figure and the more recent run rate.

The context is equally important. Stanley sits inside Stan, an established creator-commerce company that Hu says is approaching $41 million ARR. Around $38 million still comes from Stan Store.

Stan already knew creator workflows, had paying customers, had distribution and could talk directly with its target market. Hu also says the team manually simulated parts of Stanley before automating them, which helped establish that creators actually wanted the result.

The 14-day build changed the economics of experimentation. Stan could turn an idea into a paid product line extremely quickly and then keep investing once customers responded.

Stanley shows how large a vibe-coded product can become when cheap implementation meets an existing distribution machine. A random founder starting with zero audience faces a very different problem.

Does Payout prove a fully vibe-coded app can actually scale?

As seen above, Payout is currently the cleanest proof we found that an almost completely AI-built consumer app can move beyond launch revenue and reach meaningful scale.

RevenueCat documented Payout during its Shipaton competition. According to the company, Claude Code and Cursor produced the entire initial design, code and assets, with no manually written code. The app had already passed 17,000 users, 1,750 paying subscribers and $30,017 in collected revenue at that stage.

The later number makes the case much stronger. RevenueCat recently said founder Connor Burd took Payout from zero to about $80,000 MRR in four months. That corresponds to a $960,000 annualized run rate if maintained.

The growth mechanism is just as revealing. Burd spent heavily on influencer partnerships and paid acquisition. RevenueCat's later interview focused heavily on how he selected creators, measured campaigns and moved winning creative into Meta ads.

The software therefore survived the transition from an AI-built competition entry to a product being pushed through paid acquisition at scale.

Payout also gives us one of the cleanest answers to the "can vibe-coded software handle real customers?" objection. Thousands of people paid, acquisition expanded and revenue kept climbing for several months.

We still need a longer history before calling the business durable. Four months tells us far more than four days, but it does not tell us what churn looks like after two years.

Are Lovable-built apps becoming real businesses?

Yes, some Lovable-built products have already reached six-figure recurring revenue, handled real commercial workloads and even been acquired.

Lumoo is the largest clear case in Lovable's published customer stories. The AI fashion platform reported €700,000 ARR within nine months and more than 15 customer brands, including Gant, Brothers and Zoovillage. Founder Henrik Skagerlind Fasth has described the platform's functionality as 100% powered by Lovable.

Plinq followed a very different path. Founder Sabrine Matos came from growth marketing rather than engineering and built the women's safety product with Lovable. The company reported more than 10,000 users within three months and R$2.2 million ARR, approximately $450,000 at the exchange rate used in Lovable's own case study.

QuickTables offers stronger evidence of operational depth. Its founders say they built the restaurant platform in about six weeks without manually writing code. The company reached a €100,000 annual revenue pace, processed tens of thousands of transactions, grew into a larger team and was eventually acquired.

Those numbers come largely from Lovable's own case studies, so we treat them as company and founder disclosures rather than audited financial statements. The underlying businesses still leave visible footprints: named customers, user counts, transactions and an acquisition.

The broader Lovable survey puts these winners in perspective. Roughly one in five respondents reported some form of product revenue, while the majority who wanted to monetize had yet to do it. Successful examples exist, but they sit inside a huge population of unfinished, internal, experimental and unsuccessful projects.

Lovable-built product	Commercial evidence	What makes the case interesting
Lumoo	€700K ARR	Real B2B customers and enterprise fashion workflows
Plinq	R$2.2M ARR	Non-engineering founder selling a consumer safety product
QuickTables	€100K annual run rate before acquisition	Real restaurant transactions followed by an exit

Did AI coding actually create the revenue for these apps?

AI coding made the products dramatically faster to ship; the large revenue differences usually came from what founders did after the software worked.

Subscribr gives us an unusually clean demonstration. Gil Hildebrand presold 50 lifetime licenses and collected roughly $20,000 before the finished product existed. Customers had already validated the idea before AI could write the production code. He then built an audience, newsletter, SEO engine and affiliate channels around the product.

Prayer Lock tells the same story from consumer mobile. Mau Baron coded the first version in three days. He then spent months obsessing over distribution, testing short-form content across many accounts and eventually scaling paid promotion. In his Starter Story interview, he described monthly sales moving from roughly $10,000 to $15,000 and then $21,000.

ToneAdapt founder Kyan followed another distribution-heavy route. He built a very specific product for guitarists, then posted constantly on social media until a few formats started producing users. The business moved from almost nothing to roughly $25,000 in recent monthly revenue within months.

The pattern is fairly brutal: AI removed a huge amount of implementation work, which exposed the next bottleneck almost immediately.

Once a functioning app can be produced in days, customer acquisition becomes a much larger share of the job.

As seen above, Subscribr even had paying customers before the real build started. That makes it difficult to credit AI coding for creating the demand. AI dramatically reduced the cost of serving demand that the founder had already found.

Do vibe-coded ARR screenshots make the businesses look richer than they are?

Yes, early ARR screenshots can badly exaggerate how established a vibe-coded app really is, especially when a few days of unusually strong revenue are multiplied by twelve.

ARR itself is perfectly legitimate when recurring revenue has stabilized. The problem appears when an app is weeks old and people use today's revenue pace as though it will repeat unchanged for a year.

Fly is the clearest warning. Pieter Levels reached an extraordinary annualized run rate very quickly, yet he now classifies Fly among projects that made money but were not sustainable. The early sales happened. The extrapolation aged badly.

Annual subscriptions can distort the picture in a different way. A founder may collect twelve months of cash up front and then describe the current pace in annualized terms. That produces genuine cash, although the customer still has to renew a year later before we know whether the recurring business is strong.

RevenueCat's much larger app dataset shows why retention deserves attention. Its subscription research finds that only 4.6% of newly launched apps reach $10,000 in monthly revenue within two years. Reaching the number is already rare. Staying there is another filter on top.

We put more weight on businesses with several revenue checkpoints.

A young app moving through $5,000, $10,000 and $20,000 over several months gives us a trajectory. One screenshot showing a $1 million annualized pace gives us a moment.

The strongest cases now have trajectories, not just launch screenshots.

Does $20,000 a month from a vibe-coded app mean the founder is making $20,000?

No, a vibe-coded app doing $20,000 a month can leave the founder with far less once acquisition, platform fees, AI costs, refunds and operating work are included.

Stoppr is a useful case because founder David Attias discussed margins rather than stopping at the revenue headline. The app reached roughly $14,000 in monthly revenue at one stage, while Attias reported a profit margin around 20%. That works out to roughly $2,800 of monthly profit before the founder's personal taxes and before assigning a salary value to his own time.

Consumer acquisition can get expensive quickly. Prayer Lock's founder disclosed around $21,000 in monthly sales at one checkpoint, while later breakdowns placed advertising around $9,000 as he scaled. The business could still be attractive at that level, but the gap between App Store revenue and money kept by the founder becomes obvious.

AI products add another variable: inference. Every paying customer may create an ongoing model cost. A traditional small SaaS serving mostly database queries can have extremely low marginal costs. An AI product generating scripts, images or long model responses has to pay upstream providers every time customers use the expensive features.

App stores can also take a share of purchases, payment processors take another, refunds reduce realized revenue and support eventually consumes time.

As seen above, Prayer Lock reached five-figure monthly sales quickly, yet the business still had to spend heavily to keep acquisition moving. Fast coding lowers development cost. It does not automatically produce extraordinary margins.

Business	Headline revenue	Useful economic context
Stoppr	~$14K/month at disclosed peak	Founder reported roughly 20% margin
Prayer Lock	~$21K monthly sales at disclosed checkpoint	Paid acquisition became a large expense as the app scaled
AI subscription apps generally	Higher revenue per payer than non-AI apps	RevenueCat also finds higher refunds and weaker long-term retention

How rare is a vibe-coded app making $10,000 a month?

A $10,000-a-month vibe-coded app is still an exceptional result, and anyone presenting it as a normal outcome is overselling the boom.

There is no good denominator specifically for vibe-coded businesses. We can count public winners fairly easily. Counting every serious failed attempt is much harder because founders rarely publish postmortems for apps that made $87.

The broader app market gives us a useful baseline. RevenueCat analyzed more than 115,000 subscription apps covering over $16 billion of revenue. Among newly launched apps, only 17.3% reached $1,000 in monthly revenue within two years. Just 4.6% reached $10,000.

That means roughly three quarters of the apps that managed to reach $1,000 never made it to $10,000 within the measured period.

The category differences are also large. Gaming had the highest $10,000 hit rate in RevenueCat's dataset at 8.9%. Business apps were down at 1.6%.

Lovable's builder survey points in the same direction from another angle. Most respondents said they were building with revenue in mind, yet 60.5% still had no revenue and planned to monetize later. Only 10.7% reported direct product income, with another 9.1% earning from a mixture of products and services. Those groups include products earning tiny amounts, so the share reaching $10,000 must be substantially smaller.

It is a funny market today: generating an app has become normal, generating $10,000 a month from one remains rare.

The success stories are becoming real enough to study. They still sit deep in the right tail.

Has vibe coding flooded the app market faster than it created revenue?

Yes, app supply has exploded far faster than revenue has shifted toward new products, which may be the most important aggregate fact in the whole vibe-coding debate.

RevenueCat and Appfigures counted roughly 2,000 new subscription apps launching each month in early 2022. By early 2026, the number had passed 14,700 a month.

That is roughly a sevenfold increase in four years.

The acceleration became particularly sharp on iOS around the rise of AI-assisted development. RevenueCat now attributes roughly 77% of new subscription-app launches to iOS.

Yet old apps still dominate the money. Apps launched before 2020 generate 69% of the subscription revenue in RevenueCat's dataset. Apps launched in 2025 or later, the cohort closest to the vibe-coding boom, account for only 3%.

Part of that gap is obviously age. Older apps have had years to acquire customers, while a 2026 app has had months. Even with that caveat, the scale of the imbalance is hard to ignore.

The market is also getting more polarized. Median MRR growth across RevenueCat's dataset was only 5.3% year over year. The top quartile grew at least 80%, while falling more than 33% put an app in the bottom quartile.

Easier development has produced dramatically more shots on goal. Revenue is still concentrating among a relatively small set of winners.

These days, writing the software is increasingly the cheap part of launching an app. Being noticed among nearly 15,000 new subscription launches each month is much harder.

Do AI-powered vibe-coded apps keep their customers?

AI apps currently make more money from each payer early on, but their long-term retention is worse, which makes young vibe-coded revenue figures especially worth watching.

RevenueCat's latest subscription dataset gives us a surprisingly clean comparison. AI-powered apps generate 41% more realized first-year lifetime value per payer at the median: $30.16 compared with $21.37 for non-AI apps.

They also convert trials into paying customers better.

The retention numbers move the other way.

After twelve months, median monthly-plan retention is 6.1% for AI apps compared with 9.5% for non-AI apps. Annual plans retain 21.1% of AI subscribers versus 30.7% for non-AI subscriptions. Refund rates are also higher, at 4.2% versus 3.5%.

Those are meaningful gaps. An annual AI subscriber is roughly 31% less likely to remain after twelve months in the RevenueCat sample.

The data covers AI-powered apps rather than vibe-coded apps specifically, so we should not pretend it measures code quality. Many vibe-coded products sell ordinary non-AI functionality, and plenty of traditionally engineered apps sell AI.

It still identifies a problem that overlaps heavily with the current vibe-coded boom: AI makes it unusually easy to create a compelling first experience and charge for it quickly. Keeping that person paying month after month appears harder.

Current revenue trajectories deserve more weight than launch speed. A product doing $40,000 three weeks after launch is interesting. A product still doing $40,000 or more a year later tells us considerably more.

What do the vibe-coded apps making real money have in common?

The strongest vibe-coded businesses usually combine an extremely specific customer problem with much better distribution than the average builder.

Look at the products themselves.

ToneAdapt gives guitarists the settings needed to reproduce the tone of a specific song. Prayer Lock blocks distracting apps until a Christian user prays. Payout helps people find class-action settlements they may qualify for. Subscribr helps YouTube creators produce scripts. Launch Fast helps Amazon sellers research products.

Those are easy products to explain because the customer can quickly understand the desired result.

The founders also tend to bring something useful from outside programming. Hasaam Bhatti had already operated Amazon businesses before building Launch Fast. Mau Baron spent months learning mobile distribution and eventually posted huge volumes of short-form content. Gil Hildebrand built an audience before launching Subscribr. Stanley's team already served creators through a much larger business.

Non-technical founders can win for exactly that reason. Coding experience becomes less decisive when AI handles a large share of implementation, while market knowledge keeps its value.

Some newer examples reinforce the pattern. WrestleAI was built by a teenage wrestler who understood the niche and has since reported more than 100,000 downloads and close to $200,000 in cumulative revenue. Its narrow use case is immediately obvious: analyze wrestling technique using AI.

A generic AI productivity tool has to fight thousands of alternatives. A product built for guitar tones, Christian screen discipline, wrestling technique or Amazon product research starts with a clearer customer and a clearer marketing message.

ToneAdapt is now around $25,000 in recent monthly revenue. The founder's advantage was hardly an obscure technical breakthrough. He understood guitarists well enough to build something they could instantly recognize as useful.

Vibe coding seems to reward domain knowledge more than clever prompting.

So, are any vibe-coded apps still making a lot of money?

Yes. The idea that vibe-coded apps only produce disposable prototypes is unsupported now: several are making tens of thousands of dollars a month, and Stanley has reached roughly $3 million ARR.

The strongest cases cover several different markets and business models. We have a consumer app around $80,000 MRR, a creator SaaS above $60,000 a month at a reported checkpoint, Lovable-built businesses at several hundred thousand euros of ARR, and multiple solo or tiny-team apps around $20,000 to $30,000 a month.

That is already a meaningful body of evidence.

The more aggressive claim, that vibe coding has made building a profitable app easy, falls apart once we look at the denominator. RevenueCat finds that only 4.6% of new subscription apps reach $10,000 in monthly revenue within two years. New subscription-app launches have increased about sevenfold, while apps from before 2020 still collect 69% of the revenue.

We also see exactly where the difficulty moved. The profitable founders spend enormous amounts of time on paid ads, creators, SEO, partnerships, content and customer research. Some collected money before building anything. Others had an existing audience or years of domain experience.

AI coding has changed the economics of creating software much more than it has changed the economics of getting attention.

So yes, strongly. Vibe-coded apps can now become very real businesses, but the winners look much less like lucky people prompting Claude for a weekend than the hype suggests.

They ship unusually fast, then spend most of their energy doing the difficult parts AI did not remove.