Vibe-Coded Apps Making Money: 11 Case Studies, Week of Sep 28, 2026

September 29, 2026 · 1370 words

This week we published 11 case studies of apps built with AI coding tools. Reported revenue runs from $0 to $62K a month. Only one number is verified by a third party. The rest are what founders or tool makers said, so read them as claims.

AppWhatAI toolBuilderRevenue + labelRead
3AK TrackSprint and track training app (iOS)Rork; ChatGPT for debuggingNon-developer$13,343 MRR (Sep 2026) · VerifiedRead
CryptosaurusFarcaster mini app: your profile picture as a mintable dino NFTClaude, Codex, ChatGPT, GeminiUnknown (hobby coder)180+ paid $2 mints, about $360 gross · Self-reportedRead
KleoAI LinkedIn content platformClaude CodeDeveloper$62K MRR (Dec 2025) · Self-reportedRead
7-agent business experimentSeven Claude Code subagents running small online businesses for 21 daysClaude CodeUnknown$0 · Self-reportedRead
Launch FastProduct research for Amazon sellersCursor (MVP); now Codex and Claude CodeNon-developer$30K MRR (Jun 2026) · Self-reportedRead
SnapTaxTax planning for freelancers (Security Lens example)LovableNon-developer$50 MRR, 3 paying (May 2026) · Self-reportedRead
SocialCrawlOne API for public data from 67 platformsClaude CodeDeveloperAbout $12.5K/mo in credit sales (Sep 2026) · Self-reportedRead
LunairText prompt to animated explainer videoBase44Unknown$8K MRR; $100K in first 2 months (Feb 2026) · Self-reportedRead
Wrestle AIWrestling video analysis and coaching (iOS)RorkNon-developer$23K MRR (Feb 2026) · Self-reportedRead
FlagStatFlag football scorekeeping and statsBoltNon-developerNot disclosedRead
SuperXGrowth tool for X creatorsClaude Code (listed as AI-assisted dev tool)Developer$23K MRR, about 650 customers (Feb 2026) · Self-reportedRead

Vibe coded app revenue examples roundup: a grid of eleven app cards on phones and laptops, each tagged with a revenue label, one marked verified

The apps

3AK Track. Two college sprinters with no coding background built this iOS training app in Rork, an AI app builder. TrustMRR shows $13,343 in MRR (monthly recurring revenue), read straight from their billing service. That makes it the only verified number this week. Per Rork, their training brand already had 150,000+ Instagram followers. Read the 3AK Track case study.

Cryptosaurus. A hobby-level coder built this Farcaster mini app with Claude, Codex, ChatGPT and Gemini. It turns your profile picture into a dinosaur you can mint as an NFT. He reports 1,000+ users and 180+ paid $2 mints (self-reported). Launch-day load broke payments, and he refunded the users hit. Read how Cryptosaurus got its first users.

Kleo. An engineer with 10+ years of experience rebuilt this LinkedIn content tool in about four weeks with Claude Code. Kleo reported $62K MRR three months after launch, in December 2025 (self-reported, and nine months old). Two co-founders with large LinkedIn audiences, a laddered beta price and live webinars did the selling. Read the Kleo case study.

The 7-agent experiment (postmortem). A solo builder let seven Claude Code subagents run small online businesses for 21 days. Revenue: $0 (self-reported). The agents tracked their own activity, not buyers. It is an honest look at what "fully autonomous" produced. Read the postmortem.

Launch Fast. A founder who had never written code built a working MVP in 48 hours with Cursor. The tool helps Amazon sellers research products. He reports $30K MRR by June 2026 (self-reported). A partner who already had customers mattered more than the fast build. Read the Launch Fast case study.

SnapTax (Security Lens). A bookkeeper of nearly 25 years built this tax planner for freelancers in Lovable. Eight weeks in, the founder reported $50 MRR from 3 paying users out of 19 signups (self-reported, May 2026). We used it as an example of the app type. We did not test SnapTax. Read the tax-app security checklist.

SocialCrawl. A developer co-founder built this social-data API with Claude Code. It returns public data from 67 platforms in one format. They report about $12.5K a month (self-reported). Sales are one-time credit packs, so this is monthly sales, not true MRR. Read the SocialCrawl case study.

Lunair (Pricing Teardown). Lunair turns a text prompt into an animated explainer video. It was built on Base44, an AI app builder. Base44's own case study reports $100K in the first two months and $8K MRR (self-reported, vendor source). It bills per second of video, the thing that costs Lunair money. Read the Lunair pricing teardown.

Wrestle AI. A college freshman with no coding background built this wrestling-video coach in Rork in about 30 days. A Fiverr freelancer set up payments and login for about $250. Rork's case study reports $23K MRR (self-reported). Niche focus, paywall timing and influencer deals did the selling. Read the Wrestle AI case study.

FlagStat (Non-Dev Build). A self-described non-engineer built this flag football scorekeeping app with Bolt. Revenue is not disclosed. It is the official scorekeeping app of the Florida state flag football playoffs. The founder's ties to the sport won the users; the AI wrote the software. Read how FlagStat was built.

SuperX. A former software engineer built this growth tool for X creators and lists Claude Code as his AI-assisted development tool. It reported $23K MRR and about 650 paying customers in February 2026 (self-reported). How much the AI wrote is not stated. Months of posting on X came first. Read the SuperX case study.

Patterns this week

Vibe coded app revenue examples compared by band: one verified app, several self-reported $20K+ months, small earners and a $0 experiment

  1. The audience came before the app. 3AK Track had an Instagram following. Kleo had co-founders with big LinkedIn audiences. SuperX had months of posts on X. Launch Fast had a partner with customers. The AI made the build possible; the audience brought the buyers.
  2. Non-developers are in the money column. 3AK Track, Launch Fast and Wrestle AI were built by people who did not code before. Some still needed help. 3AK Track used ChatGPT to debug, and Wrestle AI paid a Fiverr freelancer about $250 for payments and login.
  3. Where the money did not come from: the build itself. The two lowest numbers this week, $0 and $50 MRR, both shipped working products. The agent experiment produced plenty of output but no buyers. SnapTax found three paying users in eight weeks.
  4. Almost every number is self-reported. Only 3AK Track is verified. Wrestle AI's and Lunair's figures come from the AI builders' own blogs, which have a reason to show wins.

Security takeaway of the week

Many AI builders store your data in Supabase, a hosted Postgres database. Your app ships with a public key to reach it. Row Level Security (RLS) is the set of rules that decides which rows that key can read. Without it, anyone holding the key can read every user's data. The SnapTax and 3AK Track case studies both point here. Turn RLS on for every table, add an owner-only policy, then test while logged out. Step by step: Supabase RLS for vibe coders.

Building something like these and want a second pair of eyes before real users arrive? Email [email protected] for a security audit of your vibe-coded app.

Key takeaway

This week, AI tools built apps in 48 hours to a few months. The money came from audiences, niches and prices the founders already understood. If you started vibe coding last month, pick your buyer before your next prompt.

Every case study so far, grouped by revenue: Vibe-coded apps making money. Turning your own project into income: From vibe code to cash.

Sources