The brief: a tracker people could actually trust
Appxcel needed a first product that proved the studio's whole approach — one app, built carefully, shipped honestly. A period and ovulation tracker was the right test case: it's used daily, the data behind it is about as personal as data gets, and most of the existing options in this category are cluttered with ads or built around monetizing exactly the data a user least wants shared. Monthlys had to do better on both fronts — genuinely useful predictions, and a private, ad-free experience.
Building predictions that adapt, instead of assuming a 28-day cycle
The easy version of a cycle tracker hardcodes a 28-day average and calls it done. That's wrong for a large share of real users, and it gets more wrong the longer someone relies on it. Monthlys instead logs a user's actual cycle history and recalculates period and fertile/ovulation windows from that history, so predictions get more accurate the longer the app is used rather than staying pinned to a textbook number.
An AI assistant that only knows what the user told it
Rather than a generic chatbot bolted on for marketing, Monthlys's AI assistant is scoped to a user's own logged history — cycle length, symptoms, patterns over time — so its answers are about that person's own data, not a canned FAQ. It's built to help someone understand their own patterns, not to replace medical advice.
Shipping on Android as a Capacitor app, not a thin wrapper
Monthlys ships as a real Capacitor Android app (package com.appxcel.monthlys) with its web assets bundled into the APK at build time, which is what makes an honest in-app "update available" check possible — the app can read its own installed version and compare it against the latest release, something a plain TWA wrapping a live URL can't do. The release APK signs inline through Gradle's own signing config, and the same web app installs on iOS as a home-screen PWA.
Distribution: the Appxcel Store, with a scan users can verify
Monthlys is distributed through the Appxcel Store rather than Google Play — the Android APK is scanned against VirusTotal before it's ever listed, giving users a verifiable clean-scan result up front instead of asking them to trust an unfamiliar download blind.
The result
Monthlys shipped as a working, adaptive cycle tracker with real update infrastructure behind it — not a prototype. It's proof that Appxcel's approach holds up end to end: product design, a working AI feature grounded in real user data, a signed and scanned Android release, and a distribution channel Appxcel controls itself. Read more about the company behind it in the Appxcel case study or the Monthlys article.