How Ritmo was built: a designer who couldn't code, Codex, Rork and a lot of run clubs

Ritmo is a dating app for people who do sport. Profiles show your sports, training times, personal bests and finished races, and the app pairs dating with local Run Clubs, so a match can turn into a run instead of another week of chat.^[1] It was built by Adel Benzehda, a French product designer living in Florianópolis, Brazil, who says he had never built a mobile or web app and didn't know how to code before it.^[1]

He has shared the build in two places: a long write-up for RevenueCat's Shipaton 2026 hackathon, and a steady run of posts on X. This piece pulls them together. Unless a line says otherwise, the numbers are his own.

The idea: a question heard at run clubs

The idea came from something he kept hearing at run clubs: "Who here is single?" People were already meeting through sport, he writes, but there was no easy way to tell who was open to something more.^[1] RevenueCat's judges opened their note on the app with the same question.^[2]

A few days before the idea, he posted a thread arguing that designers have an unfair advantage right now, because they know how a product should feel, and that one of the best things a designer can do in 2026 is build and ship their own software.^[5]

The build: idea to Google Play in 11 days

He had the idea on August 24, 2026. Ritmo went live on Google Play on September 4 and on the App Store on September 17.^[1] The Shipaton hackathon was the push, he says: without it he "might never have shipped."^[7]

He did all of it alone: strategy, branding, design, development, backend, monetization and marketing.^[1] The iOS app is native Swift and SwiftUI, the Android app is React Native with Expo, Supabase is the backend and RevenueCat handles subscriptions and one-off purchases on both stores.^[1] The "Built with" list on his entry names Codex and Rork as the AI tools, alongside GitHub and Resend.^[1] The Android package is called com.rork.ritmoandroid and its legal pages sit on a rork.app address.^[4]

He hasn't said how much of the code each tool wrote. What he has posted is a running verdict on the models: five days after the idea, "Codex with 5.6 Sol > Claude Code with Opus 5"^[6]; by October 12, "Opus 5.5 > GPT 6.1 Sol"^[10]; and on October 27, that he now uses Opus 5.5 "almost exclusively."^[19] He hasn't said whether that changed how Ritmo itself is built. Asked what one thing a new app builder would regret not knowing, his answer was blunt:

The part users never see: Ritmo Control

Next to the two apps he built a web app called Ritmo Control to run the business day to day: moderation, product analytics, revenue tracking, Run Club management and a Trello-style board for his own work. It also lets him change app settings without shipping a new release to the stores.^[1]

He also built a notification system with 14 kinds of trigger and 25 message templates in four languages. It reacts to unread messages, unanswered matches and new profiles nearby, sends evening nudges between 7 and 9 p.m. in each member's time zone, and holds back general reminders for 48 hours after a match or message alert.^[1]

How it makes money: paywalls at specific moments

Discovering people, matching and chatting are free. Money comes from weekly, monthly and six-month Pro plans plus one-off Boosts and Super Likes.^[1] Instead of one paywall, he shows offers at chosen moments:

The judges singled this out when they gave Ritmo third place in Shipaton's HAMM (Help Apps Make Money) award: they said the subscriptions, Boosts, Super Likes and contextual paywalls fit closely into the product.^[2] On October 9 he reported that about 9% of new sign-ups made a first purchase on the day they joined.^[1]

Growth: one city at a time

He launched in Brazil first, inside running and fitness communities, and backs that with localized Meta ads.^[1] On October 1 he posted on Devpost that he was "finally profitable", with an average return on ad spend of 1.2x.^[1]

A big lesson was about density. A thousand users spread across Brazil can still leave someone with almost nobody to meet nearby, so he concentrates ads on specific cities and watches the balance of men and women, the age mix (he targets an average age of 32) and whether people actually match and talk, before moving on.^[1] Most users were in five cities in southern Brazil; São Paulo is next, then Rio de Janeiro and Brasília.^[1] On X he added that 18% of paying users are women.^[13]

He says he now uses every channel he can think of, X and LinkedIn included.^[17]

The numbers so far

On October 9 his Shipaton entry listed $1,390 in monthly recurring revenue (MRR), 151 active subscriptions, 3,694 profiles, 722 matches and Run Clubs in 24 Brazilian cities.^[1] His X posts track the climb: $1K MRR on October 3^[8], from $800 on October 1 to $2,618 on October 14^[11], $4K MRR and 7,000 users on October 19^[12], and $5,500 MRR "in 40 days" (his count) on October 25.^[16]

The latest figure isn't only his word: TrustMRR, which reads revenue from Ritmo's RevenueCat account, showed $6,278 MRR from 667 active subscriptions, $6,149 in the last 30 days and $6,659 all-time on October 27, 2026.^[3] AppBrain, an independent app-data site, estimates about 2,700 Android installs, with a best rank of #88 among free dating apps in Brazil.^[4]

Timeline

1. Aug 20, 2026

   Posts that designers should build and ship their own software.^[5]

2. Aug 24

   The idea.^[1]

3. Aug 29

   "Codex with 5.6 Sol > Claude Code with Opus 5."^[6]

4. Sep 4

   Live on Google Play.^[1]

5. Sep 17

   Live on the App Store.^[1]

6. Oct 1

   He posts "$10k ARR" and, on Devpost, that he's profitable.^[7][1]

7. Oct 9

   $1,390 MRR and 151 subscriptions in the Shipaton entry.^[1]

8. Oct 14

   $2,618 MRR.^[11]

9. Oct 21

   Third place, HAMM Award, Shipaton 2026.^[2][15]

10. Oct 25

    $5,500 MRR.^[16]

11. Oct 27

    $6,278 MRR, 667 subscriptions (TrustMRR, RevenueCat-verified).^[3]

Lessons for builders

1. 01

   Use the skill you already have

   He argues designers have an edge because they know how a product should feel; his favourite detail is the 3D streak bottle he pointed the judges to.^[5][1]
2. 02

   Ship to one store first

   Android went live in 11 days; iOS followed 13 days later.^[1]
3. 03

   A deadline helps

   He credits the hackathon for getting him to ship at all.^[7]
4. 04

   Build your own control panel

   A small admin app for moderation, revenue and remote settings means you don't need a new store release for every change.^[1]
5. 05

   Ask for money at the right moment

   Offers tied to moments (after onboarding, after a missed match) rather than one wall at the door.^[1][2]
6. 06

   Win one city before the next

   For anything that needs people nearby, density beats total sign-ups.^[1]
7. 07

   Keep testing the models

   His favourite model flipped within two months, so check again before you settle.^[6][10][19]