Steal this idea: Tarsi's paid-upfront launch up the empty paid charts

Tarsi is an offline-first budget tracker for iPhone and Android made by Bryl Lim in the Philippines. He built it as a live demo for a university talk about vibe coding: instead of a slide deck, he showed a real app. He told the Tap & Swipe podcast he primarily used Codex, on top of React Native templates he already had, and made the tarsier mascot with Gemini's Nano Banana from a Bulbasaur reference image.^[1] Then he published it, and people started buying it.

The interesting part isn't the weekend build. It's how he charged for it. Most new apps launch free with a trial and a paywall. Tarsi launched as a paid download, no trial, and that one choice is what this page is about: what the pattern is, what it did, how he repeated it, and what changed when Tarsi itself went free in September 2026.

The pattern in one sentence

Ship a small, complete app as a one-time paid download, in a category where the paid chart is thin, so a modest number of daily sales puts you at the top of that chart, and let the rank and your community do the marketing.

What it did for Tarsi

Apple lists Tarsi's release as 12 March 2026.^[9] The next day Bryl posted: "I posted my app 5 hours ago — now it's the #1 Finance app on the App Store".^[2] On the podcast he read out about $14.9K for iOS from App Store Connect, and "around 19 or 20" thousand including Google Play, which had only been live for about three weeks.^[1] Tap & Swipe's episode page adds about $350 a day in sales on average (up to $800 on days it trended on social media) from 80-90 downloads a day, which it says was enough for #1 on the paid charts in several countries.^[1] In August he wrote that Tarsi had been "the official #1 Finance app on the App Store and in 13 other countries for 6 months straight".^[4]

1. Mar 2026

   Built in a weekend with Codex for a university talk; released on the App Store on 12 March.^[1][9]

2. 13 Mar 2026

   #1 Finance app on the App Store five hours after he posted it (founder-reported).^[2]

3. 7 Apr 2026

   Tarsi launches on Google Play.^[3]

4. Apr 2026

   Same playbook for a trip planner (Mayi) and a workout tracker (Kabi), both paid.^[13][14]

5. 23 May 2026

   Tap & Swipe episode: ~$20K first month, no paid ads (founder-reported).^[1]

6. Aug 2026

   "6 months straight" at #1 in Finance, by his count.^[4]

7. 25 Sep 2026

   Version 2.4.1: "Tarsi is now FREE!", with Pro as an optional upgrade.^[9]

8. 5 Oct 2026

   A new paid finance app, Zero, goes live from the same developer.^[15]

Why it worked, in his words

It was the fastest way to ship. Asked why he went paid upfront, his first answer was "The truth is, it was easier that way." No paywall setup, no RevenueCat or other payment integration on his side; the App Store handles the money. He says the main reason was speed.^[1]

Buyers are tired of subscriptions. He found people were willing to pay with no free trial, and thinks that in a market full of freemium apps, an app that gives you everything without a subscription feels like "a breath of fresh air".^[1]

The paid charts are less crowded. When someone asked what went right after his workout app hit the top of the charts in 16 countries, his whole answer was:^[5]

No server, no running costs. The first version had no database and ran on the device, so each sale cost him nothing to serve. Server-backed features (sync across devices, AI features, bill splitting) came later as optional subscriptions, because those cost him money to run.^[1]

He kept running it

On the podcast he said his trip planner followed the exact same playbook and became the number one paid travel app, "which meant my strategy worked"; he brought in friends for equity to run the community and features because he didn't have the bandwidth for two apps.^[1] His App Store developer page now lists 11 apps, and 9 of them are paid downloads.^[8] Here are the ones in the story so far, as of 29 October 2026:

Chart positions move by the hour, so treat these as a snapshot. But notice the shape: the three older paid apps sit near the top of their category charts, and the one that went free sits much further down its category list. The October batch looks like the same bet spread across six quiet categories at once; it's too early to say how those are doing, and we haven't seen him post numbers for them.

The twist: Tarsi went free

On 25 September 2026, version 2.4.1's release notes said "Tarsi is now FREE!", with Pro as an optional upgrade; two days later 2.4.2 added Tarsi Pro Lifetime, "pay once and keep every Pro feature forever".^[9] Tarsi's pricing page now says the free tier covers logging income and expenses, up to 2 accounts, 1 category budget and two months of history. Pro is ₱199 a month or ₱1,999 a year (both with Cloud sync), or ₱1,999 once for life without Cloud. And for earlier buyers: "Bought Tarsi before it went free? You keep Pro for life".^[10]

He hasn't explained the switch anywhere we could find. Ten days before it, he asked his followers on X: "Is it better for an app to have a large number of users with low profit per user, or a smaller user base with high profit per user?"^[7] We don't know his answer, so we're not guessing at it.

What we can see: TrustMRR, which reads Tarsi's RevenueCat data, shows $971 in monthly recurring revenue from 373 active subscriptions and $2,212 in the last 30 days, as of 29 October 2026.^[11] That covers the in-app purchases only, not the earlier paid downloads (Shipped & Paid notes the RevenueCat data starts with the subscription plans in April 2026).^[12] And his newest finance app, Zero, launched ten days after Tarsi went free as a pay-once app again.^[15] So the paid-upfront launch isn't a forever model for him. It looks more like a way to start.

How to run it yourself

Most of these steps are Bryl's; the first and last are our suggestions, built on what his story shows.

1. 1

   Look at the paid chart before you pick the idea

   Our suggestion: open the top paid list for your category in your country and a few others, and see how much is there. Bryl's whole edge was that the paid side had fewer competitors.^[5] On the episode page, Tap & Swipe says 80-90 downloads a day was enough for #1 in several countries.^[1]
2. 2

   Build a small, complete app fast

   One weekend, Codex, and templates he already trusted.^[1] A paid app has no trial to win people over, so what you sell has to work on day one.
3. 3

   Keep it on the device

   No database, no account, no server at launch, so every sale is margin and nothing breaks when a chart spike hits.^[1]
4. 4

   Charge once, and make the store page the whole promise

   Everything in his App Store description and screenshots comes with the purchase. He didn't put the later cloud features in the screenshots at all, because that "might be misleading".^[1] He made the screenshots in Figma with a free mockup plugin, used ChatGPT to draft the headings, used all 10 screenshot slots and focused on the first three.^[1]
5. 5

   Launch it from your own channels

   No paid ads. He showed it at the talk, posted it on X, and otherwise used the channels he already had, mostly Facebook and Threads.^[1][2]
6. 6

   Put a public community link in the app

   A link in Tarsi's settings goes to a public Facebook group for questions and suggestions; he points feedback to App Store reviews instead of a form. His reasoning: activity in a public group shows up in members' friends' feeds, which is free marketing.^[1] He also tested a joke feature there (sharing your daily spending to Instagram Stories, like Strava) and shipped it with the app name as a watermark.^[1]
7. 7

   Collect Android buyers before you build for Android

   He had no Android plans until a simple Google Form waitlist got 900 beta sign-ups. Some testers bought the app during the beta.^[1] If your Google Play personal account was created after 13 November 2023, you need a closed test with at least 12 testers opted in for 14 days in a row before you can apply for production, so a waitlist helps here too.^[16] He priced Android $2 lower because it has fewer features (no receipt scanning or Apple Intelligence), not because of the platform.^[1]
8. 8

   Add subscriptions only for what costs you money

   Cloud sync, bill splitting and AI features run on servers, so those became optional subscriptions, framed as "optional quality of life features".^[1] Tap & Swipe's write-up says there had been zero complaints so far.^[1]
9. 9

   If you switch to free later, look after the people who paid

   Our suggestion, from what Tarsi did: when it went free, earlier buyers kept Pro for life.^[10]

Gotchas before you copy it

• No recurring revenue. The podcast host put it plainly: paid upfront means you don't have recurring revenue.^[1] Every month starts from zero, so the rank has to keep selling the app.
• #1 in a thin chart is a small number. Tap & Swipe's figure is 80-90 downloads a day.^[1] That's the point of the pattern, but it also means the ceiling is lower than it looks from the badge.
• You will be told you shipped too early. He said he got "a lot of flak" from other developers for publishing without proper testing; his view is that AI lets him iterate fast on feedback and validate the idea first.^[1]
• Demand can outrun you. In September he wrote that demand for Tarsi was "really unprecedented" and that he might have to pause new downloads for a bit to keep service good for existing users.^[6]
• He had a head start. Shipped & Paid notes his X bio calls him a "300k+ creator" (with about 1.8K followers on X itself), and the first users came from his talk and his socials.^[12][1] A paid launch with no audience at all will start slower.

What to take from this

1. 01

   Pick the less crowded chart

   Paid had fewer competitors, in his words.^[5]
2. 02

   Ship the simplest money path

   Paid upfront meant no paywall to build.^[1]
3. 03

   No server until it pays

   Offline first; cloud later, as an upgrade.^[1]
4. 04

   Sell exactly what's on the page

   Screenshots show only what the price buys.^[1]
5. 05

   Make the community public

   He says group posts reach members' friends.^[1]
6. 06

   Treat the model as a phase

   Tarsi went free; Zero launched paid.^[9][15]