Extra messages unlock for the day
Every category ships with sixty messages that are always free. Watching an ad adds bonus messages on top, and the unlock expires with the day rather than becoming a permanent entitlement to track.
Bytecraft product · Android · Consumer · Ad-supported
Romantic Love Messages: an app built around sending, not scrolling.
A consumer Android app for finding a message, making it your own, and sending it to one particular person. Everything technical in it exists to serve that last step.
The product problem
A message app's real job starts after the message is found.
There is no shortage of apps that show you romantic quotes, and almost all of them stop there — a list to scroll, an ad between screens, nothing that leaves the app. But nobody opens this kind of app to read. They open it because they want to say something to someone and cannot find the words. The product only succeeds at the moment something is sent.
-
Find something
Curated categories, a quote each day, and a random draw for the people who do not know what they are looking for.
-
Make it theirs
Every message can be edited, restyled and turned into an image. A message sent word-for-word is the weakest version of this product.
-
Send it
Copy, share sheet, or a saved image. The app is designed to be left — and to be worth opening again tomorrow because of it.
The core screen
Five things you can do with one message.
Every card in the app carries the same five actions. Deciding what they were — and that none of them would be held back — set the shape of everything built afterwards.
- Edit
- Opens the message in the composer with the text already in it. Nothing written by us is treated as final.
- Listen
- Reads the message aloud. It helps anyone who finds a wall of text hard to scan, and it is a genuinely different way to receive something written to be felt.
- Copy
- The plain-text exit. Most messages leave this app for a chat thread somewhere else, and pretending otherwise would only have made the product worse.
- Share
- Hands the message — or the composed image — to the Android share sheet, so the user picks the destination rather than us.
- Like
- A private save. Liked messages are the only signal the product has about what a specific person actually finds worth sending.
The hard part
The composer, where a message becomes something you can send.
Writing text over a colour is easy. Writing text over a photograph somebody else chose, at a size that stays legible after a messaging app has compressed it, is not. The composer is where most of the engineering in this product went.
What the composer had to guarantee
- Text stays legible on backgrounds we did not choose.
- The exported image matches what was on screen.
- Long messages reflow instead of overflowing the canvas.
- Every export carries a small credit back to the app.
- Saving, copying and sharing all produce the same result.
The five-hundred character limit exists to protect the last two.
Discovery
Three ways in, for three different people.
Search only helps someone who already knows what they want, and in this product almost nobody does. So the app opens with three different answers to "where do I start".
- Categories For someone who knows the occasion — an anniversary, an apology, a first message. Each category is a curated set rather than a generated one.
- Quote of the Day A single message, changed daily. It gives the app a reason to be opened that does not depend on the user having an idea first.
- Surprise Me One message at a time, drawn at random, with the same five actions attached. For the largest group of all: people who want to be handed something.
Monetization
How it earns without putting the product behind a wall.
This audience will not pay a subscription for messages, and an app that interrupts someone mid-sentence to show an ad gets uninstalled before it earns anything. Both constraints were known before the first screen was designed, so the access model was built into the product rather than fitted around it afterwards.
Premium backgrounds unlock the same way
The composer's locked gradients and photo backgrounds open for a day after a short ad. The free styles are enough to make something worth sending, so the ad buys a nicer result — never the ability to send at all.
Ad demand comes from more than one network
A single ad network fills unevenly by country, and this audience is spread across many. Letting several networks compete for the same slot matters far more to revenue than adding more ad slots ever would.
The rule underneath all three: an ad may buy something better, never the ability to use the app at all.
In production
Shipped, monetized, and quiet.
The app has been live on Google Play since launch, with advertising enabled and updates shipped against real usage. It has never been removed or suspended, and the crash and ANR rate has stayed low enough that reliability work has never displaced product work.
We publish stability rather than download counts. Installs describe a store listing; stability describes what we built.
- Crash and ANR rate
- Under 0.5%
- of installs affected across the first two months on Google Play.
- Store record
- No removals
- live continuously since release, through policy and target-API updates.
- Revenue model
- Ads only
- rewarded and mediated, with no paywall on the core content.
Decisions
The questions that shaped the product.
The finished screens are the least interesting part of any case study. These are the forks — the points where the app could have become something else.
- Should the app try to keep people inside it?
- No. The whole point is to send something to another person, so Copy and Share sit on every card. Optimising for time-in-app would have pointed this product in exactly the wrong direction.
- What is actually shipped when someone taps Share?
- Either plain text or a rendered image, depending on where they came from. The composer produces a real image file, which is why background, typography and legibility became engineering problems rather than design preferences.
- How do you keep text readable over an arbitrary photo?
- You cannot solve it once. A fixed shadow fails on a bright sky and a fixed dark overlay ruins a dark portrait, so the composer hands the user a readability control and lets them settle it against the image in front of them.
- Should ads interrupt, or should they be offered?
- Offered. Every ad in the product is one the user chose to watch in exchange for something specific. That produces fewer impressions than forced interstitials would, and far fewer uninstalls.
- Why watermark the composed image?
- Because a message app's best distribution is the message. A small credit on an exported image is the only growth mechanism that travels with the thing people were already going to send.
- Is a five-hundred character limit a restriction?
- It is a design decision. Past roughly that length the text stops fitting a shareable image at a readable size, so the limit protects the output rather than the storage.
Technology
What this app is built with.
A consumer app that earns from advertising lives or dies on three things: how fast it opens, how well the ad stack fills, and whether you can see what changed after an update.
- Kotlin
- Java
- Swift
- AdMob
- AppLovin
- Liftoff
- Meta Audience Network
- Firebase Analytics
- Google Analytics 4
- PostHog
- Microsoft Clarity
- Sentry
Shipped
What is in the released product.
- Curated message categories
- Quote of the Day
- Randomised discovery
- Text-to-speech playback
- Image composer
- Gradient and photo backgrounds
- Readability control
- Image export
- Android share-sheet integration
- Favourites
- Rewarded-ad unlocks
- Ad mediation
- Analytics and crash reporting
- Google Play release and updates
Why this project represents how we work.
A consumer app with no subscription has nowhere to hide. It has to be worth opening without a reminder, worth using without an account, and profitable without charging anyone — which means product, engineering and monetization cannot be three separate conversations. Running our own app in that market is how we learned which advice is actually worth giving on someone else's.
Building a consumer app that has to pay for itself?
Tell us who it is for and how you expect it to earn. We will be straight with you about whether the model works before we talk about screens.
Rather talk it through? Book a discovery call.