Almost every first meeting opens with the same question: "What does it cost to get an app built?"
I don't answer it with a number. Not because there's something to hide. A figure without your scope is almost always wrong, and once it's been said, it sticks in every budget meeting after. Two apps that sound the same in one sentence can be months apart once we've seen the integrations, the data and the people who'll use them.
So with us it goes like this. First comes a free 4-hour workshop. We go through your requirement with you, ask the awkward questions and suggest the solution we'd build ourselves. Then you get a written plan and the cost: once the scope is clear, the quote is with you within three working days, as a fixed price.
What I can give you before that, in thirty seconds, is the order of magnitude in team and weeks, plus the real cost drivers. Here are both, along with everything else I normally explain across the table.
The numbers
A first production-ready release — architecture, design, development, QA, release, everything included — takes one of these four shapes with us:
| Project shape | Team | Duration |
|---|---|---|
| Lean MVP — one core journey, in front of real users fast | 2–3 engineers | 8–14 weeks |
| Standard product — several roles, integrations, admin surface | 3–4 engineers + delivery lead | 12–20 weeks |
| Complex platform — multi-tenant SaaS, hardware, regulated data | 5+ engineers | 20+ weeks |
| AI voice agent, v1 — telephony, qualification, human handoff | AI engineer + backend | 10–20 weeks |
Why no amounts? Because one row can hold a project with one integration and a project with five, and those sit far apart on price. Swiss clients get their quote in francs, at a fixed price — the number you sign is in your currency. For typical market ranges there's our Swiss cost guide.
And so we understand each other: team sizes and durations are indicative 2026 figures, not a quote. What moves the number is complexity — integrations, compliance, data migration and the reliability bar you need on day one. Your fixed price comes in writing after the workshop. If the scope is still open by then, a short paid discovery sprint comes first and is credited in full against the build.
Where the numbers come from
No magic. A working month with us is 160 hours, four 40-hour weeks. Every feature block costs engineer-weeks times the monthly rate of the role doing the work. That's the whole formula.
What's interesting is which blocks are expensive. Login, lists, a feed — routine. But a cart with payment processing is a different animal: a payment bug costs real money, so it gets built differently, tested differently, secured differently. Same for offline sync, chat, live tracking, media streaming. If one of those words is in your requirements list, you're several engineer-weeks further along.
Then the platform question. We build iOS and Android from one codebase; the second store release still adds roughly a quarter of the feature build on top — store compliance and device QA don't disappear just because the code is shared. A web surface on top: about the same again. If the web surface is the actual product rather than an addition, it runs on its own calculation. What a web app costs and how the project runs is written up separately.
Design: our existing design system, adapted to your brand, is the fastest route to launch. Your own design language with research and prototypes — nice, sometimes right, about +20%.
And compliance. Data-protection basics (GDPR, in Switzerland the nDSG) are part of our standard setup — no surprise surcharge. Regulated is something else: health, finance, public sector. That gets built in from sprint one — retrofitted, it's not an update, it's a rebuild.
If you'd rather play this through yourself: our cost calculator shows the engineer-weeks, the role and the arithmetic behind every block. No email address required. That mattered to me — I hate calculators that turn out to be a form.
Three projects to measure against
Numbers without context stay abstract. So: three projects we built and still operate today. What the clients paid, I won't write — we don't publish client numbers, full stop. But you can place the project shape exactly.
FeelEat
FeelEat sells fresh meals through connected fridges, across Switzerland. Sounds like "an app". It isn't. It's scan-to-unlock on physical devices, in-app payment, live inventory, temperature telemetry, kiosk interfaces, an admin portal for operations.
That's the "complex platform" category, no debate: hardware, money, several surfaces, continuous operation. The same team has shipped every release for over nine years. Its fridges lose roughly 90% less product to theft than honour-system fridges. The whole story is in the portfolio.
As soon as physical devices, money or several user groups are involved, you're in the upper category. That stings in the proposal and saves you in operations.
LeadTrack AI
For an Australian client we built LeadTrack AI: a multi-tenant SaaS platform whose AI voice agents call and qualify inbound leads before a human picks up — first call in under 30 seconds, while the lead is still at their screen. Over 100,000 calls in production, +38% conversion. Case study.
That's the fourth row in the table above, the AI voice agent. I kept it as its own category deliberately: telephony, real-time speech and human handoff are their own discipline — but the entry point sits below a complex platform. If you're currently thinking about an AI phone assistant for your business: this is exactly the stack behind it, and at the end you own it yourself instead of renting it on a subscription.
HappyBee
HappyBee, from the FeelEat family, shows parents what their children eat at school: weekly menus, nutrition, per child. One purpose, one user group. The shape of a standard product — and honestly the shape most ventures should start in. Case study here.
A product that does one thing right. Extending comes later, on a clean base. The other way round works too — it's just more expensive.
What's not in the quote
Three items that come on top with every vendor, and that remarkably little gets written about:
Cloud and licence fees. AWS, Azure or Google bills you those directly — or we pass them through at cost with the receipt attached. No margin, but on top.
VAT and local taxes.
And operations. Our rule of thumb: annual support typically runs at 15–20% of the original build cost. Our care plans are sized in hours — 15, 30 or 60 a month, depending on how much the product is still moving. The 30 days of free fixes come first anyway. From four people on the team there's also a project-management line in the proposal. In the proposal, not in the fine print — that's the difference to look for with any vendor.
My actual advice
Don't buy a big estimate. Buy a small finding.
If your scope is still open, our discovery sprint is that finding: one to two weeks, paid. At the end you have a written scope, an architecture, a costed roadmap and an explicit "we won't build" list. The document is yours, even if you build with someone else afterwards. Build with us and the sprint is credited in full.
Before that comes the workshop: four hours, free. If the scope is clear after it, you have a written scope and a fixed-price quote within three working days, no sprint needed.
First know what's being built. Then a number that holds. The order sounds banal, and yet half the industry runs it the other way round.
Frequently asked questions
What does a simple app cost? A lean MVP — one core journey, iOS and Android from one codebase — takes 2–3 engineers and 8–14 weeks with us. Anything offered suspiciously cheap almost always saves on QA, operations or code ownership. The price hides somewhere.
What does an app cost per month, after launch? Budget 15–20% of the build cost per year for support, updates and monitoring — divide by twelve and that's your monthly line, plus your cloud bill.
Why are there no CHF prices on your website? Because the right number depends on your scope, and I'd rather show no number than one that doesn't fit your project. Your quote comes in CHF, in writing, within three working days of the workshop, at a fixed price when the scope is clearly defined. Typical market ranges for context: the Swiss guide.
Development in India — does that work seriously? With the right model: German-speaking contacts in Switzerland, deployment on infrastructure in your jurisdiction, NDA before any technical discussion, data processing agreement, 100% of code and IP with you. That's how we work — the model in detail. On experiences and risks I promised a separate, unvarnished article here; this is it: getting an app built in India.
How accurate is your online calculator? As accurate as it can be before the workshop: it shows team, weeks and phases for every block, every line visible, but no price. It doesn't know your legacy systems. An order of magnitude for planning, not a quote. Here it is.
If we're not the right team for your project, I'd rather tell you in the workshop than after weeks of proposal meetings.
If you're crunching numbers right now: the calculator takes two minutes. If you'd rather talk: the first workshop costs nothing.
