Apps with real users, not pretty prototypes
iOS and Android through publication in both stores. The last one we built and maintained has more than two million active users across ten countries.
Key takeaways
- Building the app is half the work: the other half is store review, developer accounts, signing certificates and annual maintenance.
- An unmaintained app stops working on its own within about two years, because iOS and Android ship yearly versions that break things.
- Quarl uses React Native for most cases — one codebase for iOS and Android — and native only when the case genuinely demands it.
- If the idea is not yet validated with real users, an app is a very expensive way to find out: a mobile web experience or a WhatsApp flow costs a fraction.
Operating track record
Three years operating the system in production — not one delivery and an exit.
The most frequent failure in RAG architectures, solved with structured metadata at ingestion.
The proposal arrives within 48 hours with fixed scope, price and date. If scope changes, it is quoted separately and approved first.
The app is half the work
The other half is what happens after it is ready, and that is where most mobile projects that end badly get stuck.
- Store review. Apple rejects for reasons that are on no obvious checklist, and each rejection costs days. You need to know what they check and prepare the submission for it.
- Developer accounts. Apple charges USD 99 a year, Google USD 25 once. They must be in your company’s name: if they belong to whoever built the app, the app is not entirely yours.
- OS updates. Every year iOS and Android change things that break apps. An unmaintained app stops working on its own within about two years.
- Signing certificates. Losing the Android signing key means never being able to update the app again. They are handed to your company and documented.
One codebase, two platforms
React Native for most cases: a single codebase for iOS and Android, which is the difference between one project and two. Native only when the case genuinely demands it — heavy graphics, intensive sensor use, very specific system integrations — and we justify it before proposing it.
And mobile almost never comes alone: it needs backend, authentication, push notifications, payments or integration with what you already have. We build that too, so you are not coordinating two vendors who blame each other when something breaks.
What a mobile app costs
Nobody publishes this. Here are honest ranges: price is driven by how many features it has, how many systems it connects to, and whether it needs its own backend.
| Type | Example | |
|---|---|---|
| Catalog or content app | Information lookup, no transactions or complex accounts. | 6–8 weeks |
| Transactional app | User accounts, payments, notifications, own backend. | 10–16 weeks |
| Full platform | App, admin dashboard and integrations with existing systems. | 4–6 months |
Recurring costs on top: Apple USD 99/year, Google USD 25 once, plus infrastructure. Annual maintenance usually runs 15% to 20% of build cost.
If your idea is not yet validated with real users, an app is a very expensive way to find out.
In many cases it is better to start with a mobile web experience or even a WhatsApp flow, confirm people use it, and build the app once you know which features matter. We will say so on the first call if that is your case.
Related services
Software development
The system around the model, not just the model.
View serviceIntegrations and APIs
MCP servers against the systems already in operation.
View serviceChatbots and assistants
With the evaluation set run before it ever reaches production.
View serviceAI consulting
Architecture, evaluation criteria and cost per query before writing code.
View serviceFrequently asked questions
React Native or native?
React Native for most cases, because a single codebase for iOS and Android is the difference between one project and two, in cost and in maintenance. Native when the case demands it: heavy graphics, intensive sensor use, very specific OS integrations or extreme performance requirements. We justify it with concrete reasons before proposing it.
How long until we are in the stores?
Build time depends on scope, six weeks to six months. App Store review usually takes one to three days, but a rejection adds days and sometimes weeks if it forces product changes. Google Play is faster. We prepare submissions for Apple’s review criteria, which is where most projects get stuck.
Whose name are the developer accounts in?
Your company’s, always. It seems like a minor detail and it is not: if the accounts belong to the vendor, you cannot change vendors without losing the app and its reviews. We set the accounts up in your name from the start and hand over documented access and certificates.
What about maintenance?
Every year iOS and Android ship versions that break things, and the stores raise their minimum requirements. An unmaintained app stops working on its own within about two years. Annual maintenance usually runs 15% to 20% of build cost, and it is optional: you can do it with your team if we document well, which is what we do.
Do you include the backend?
Yes. An app almost never lives alone: it needs authentication, an API, a database, push notifications and sometimes payments. We build it end to end so you are not coordinating two vendors who blame each other when something breaks.
Can you add AI features to our app?
Yes, and it is one of the highest-value combinations we see: an in-app assistant with the user’s context and permissions, semantic search over your content, or data extraction from the camera. We apply the same measurement discipline as in any AI project.
What if our idea is not validated yet?
We will suggest not building the app yet. Validating with a mobile web experience or even a WhatsApp flow costs a fraction and tells you which features actually matter. Building the app afterwards, knowing that, is cheaper than building it twice.
Book 15 minutes
Tell us what you are building, or what stopped working. You leave the call with a concrete answer: it can be fixed, it can be built, or it isn't worth it.