Building an app in 2026 costs anywhere from €15,000 for a focused MVP to €350,000 or more for a complex, regulated product. Most real business apps — the ones with logins, payments, and a backend — land between €40,000 and €150,000. The honest answer to "how much does it cost to build an app" is a range, and the only thing that narrows it is knowing how complex the app actually is and who builds it.
Everyone else opens with "it depends." That's true and useless. Below we make it depend on things you can actually decide: the complexity of the app, the features inside it, and the build method. Fix those three and the number stops being a mystery.
One distinction up front, because it changes everything. A website is content you read. An app is software you use — accounts, a database, logic that changes based on who's logged in. If what you need is mostly pages, read how much a website costs instead. This post is about real software.
Here is the realistic 2026 spread for building an app, before recurring costs, from a competent team in Western Europe:
What moves the number, in order of impact:
1. How many features, and how custom. A to-do list and a ride-hailing platform are both "apps." The gap between them is 20x or more. Every screen, state, and edge case is build time.
2. The backend. The part founders underestimate hardest. Auth, payments, a database, realtime sync, an admin panel — none of it is visible in the app, and all of it is real engineering. A "simple" app with a serious backend is not a simple app.
3. Platforms. iOS only, Android only, or both. And whether you build native (two codebases) or cross-platform (one codebase, React Native or Flutter). This single decision can swing the budget 35–50%.
4. Who builds it. A no-code tool, a freelancer, an offshore agency, a boutique studio, and an in-house team are five different products — not five quality grades of the same thing. Each carries a different price and a different risk profile.
This is the table most people actually want. Ranges are for the build only — design, development, QA, and initial store deployment — in 2026 euros, assuming a competent freelancer or small studio at Western European rates. Not no-code, not a large consultancy. Adjust using the build-method section below.
| App complexity | What it is | Typical build cost | Build time |
|---|---|---|---|
| MVP | One platform, 3–5 core screens, thin backend, validate the idea | €15,000–€40,000 | 6–12 weeks |
| Simple app | One platform, basic features, light backend, no payments | €20,000–€50,000 | 2–4 months |
| Medium app | iOS + Android, accounts, payments, push, a few integrations | €50,000–€120,000 | 4–8 months |
| Complex app | Realtime, marketplace logic, heavy integrations, custom backend | €120,000–€250,000 | 7–12 months |
| Enterprise / regulated | Fintech, health, compliance, scale, security audits | €250,000–€350,000+ | 9–18 months |
A few notes the table can't carry:
MVP. The honest truth: the cheapest version of a good MVP is not a cheap app — it's a small app built well. The cost comes down by cutting scope, not by cutting quality. One platform, the two or three features that prove the idea, and a backend that does only what those features need. We'll make the full argument for this below, because it's the single best lever you have on cost.
Simple app. A genuinely simple app — a calculator, a habit tracker, a single-purpose utility — really can land near the bottom of the range. The moment it needs user accounts that sync across devices, it stops being simple, because now there's a backend.
Medium app. This is the most common "real" product, and the widest part of the market. The range is wide because "payments and accounts" hides a lot: one Stripe integration is a week; a multi-currency, multi-provider, subscription-with-proration billing system is two months.
Complex app. The range is brutal because complexity compounds. Realtime features (live chat, collaborative editing, location tracking) need infrastructure that never shows up in the UI. Marketplaces are really two or three apps — buyer, seller, admin — wearing one name.
Enterprise / regulated. Here the software is often the smaller cost. Security audits, penetration testing, GDPR and sector compliance (HIPAA-equivalent, PSD2, KYC), and the documentation regulators demand can match or exceed the build itself.
The complexity tier is a starting point. These are the specific things that move you up or down inside it.
The single biggest driver, and the one founders control most directly. Cost scales with screens, states, and edge cases — not with your idea's ambition. "Add a chat feature" sounds like one line in a brief. In reality it's message threads, read receipts, typing indicators, push notifications, offline handling, moderation, and a backend that stores and routes it all. Every feature has an iceberg under it.
The part you can't see and can't skip on a serious app. Rough 2026 add-ons on top of the front-end build:
Here's our actual philosophy on this, and it shapes every quote we write: backend gets added when you actually need it, not by default. Plenty of MVPs ship with a deliberately thin backend and a manual process behind the scenes, and that's the right call at the start. You build the expensive infrastructure once the usage justifies it — not on day one because a template had it.
Maps, analytics, messaging, calendars, CRMs, ERPs, payment providers, AI APIs. Each integration is €1,500–€8,000 of work depending on how cooperative the other side's API is — and some of them carry their own monthly fees forever (see recurring costs below). Three integrations is a normal medium app. Ten integrations is a different project.
Not decoration — it's how the app gets used. A polished, custom-designed app costs €8,000–€40,000 in design alone for a medium product, covering user flows, a design system, prototypes, and the dozens of empty/loading/error states that separate a real app from a demo. Skimping here is the most visible way an app looks cheap.
The decision with the biggest single leverage on cost. Building two native apps (Swift for iOS, Kotlin for Android) means two codebases, two teams, two timelines — and roughly €40,000–€110,000 for a dual-platform medium app. The same app in React Native or Flutter — one codebase, both platforms — typically runs €15,000–€45,000 for the equivalent scope. That's a 35–50% saving, and for the vast majority of business apps the user can't tell the difference.
Native earns its keep for graphics-heavy games, deep hardware integration, and apps where a few milliseconds of UI performance is the product. For everything else — which is most apps — cross-platform is the sensible default in 2026. We made the full case in React Native vs. Flutter for startups.
Shipping both roughly adds 30–40% to a native build (two of everything). With cross-platform, the second platform is far cheaper — but never free. Store-specific behaviour, device testing, and platform review quirks still cost real time on both sides.
Cheap below a certain bar, expensive above it. A standard app with sensible practices: baked into the build. A regulated app needing audits, encryption-at-rest, penetration testing, and compliance certification: €20,000–€100,000+ on top, sometimes recurring annually. If you handle money, health data, or children's data, this is not optional and not where you economise.
Same app, five ways to build it, five very different invoices — and five very different risk profiles.
You assemble the app in a visual tool instead of writing code. Realistic 2026 cost: €2,000–€20,000 to build, plus €30–€500/month in platform fees forever.
The honest take on no-code: it's not a toy, and dismissing it is lazy. For the right job it's the smartest money you'll spend. But if the app succeeds, you'll likely rebuild it in real code — so treat the no-code version as a paid experiment, not a foundation. The trap is building your future on top of a tool you'll outgrow.
A skilled freelancer at €50–€90/hour in Europe (less from lower-cost regions, more for specialists). A simple-to-medium app: €15,000–€60,000.
Teams in lower-cost regions billing €25–€70/hour. The headline saving is real — but so is the overhead.
This is us, and the honest pitch is: you pay for senior judgement, not headcount. A small, senior-only team — no juniors, no account managers — builds the app and talks to you directly. Western European rates, but no overhead tax. A medium app: €50,000–€120,000.
Your own developers. The most expensive path up front — a single senior mobile engineer in Western Europe is €70,000–€120,000/year fully loaded, before design, backend, and QA roles.
The build is a one-time number. Running an app is a forever number — and it's where "cheap" quotes get their revenge. Real 2026 figures:
Maintenance. The big one. Budget 15–20% of the build cost per year just to keep the app working — bug fixes, dependency updates, and the OS changes Apple and Google force on you twice a year. A €60,000 app costs roughly €9,000–€12,000/year to maintain before any new features. An app you don't maintain stops working within a year or two; this is not optional.
App store fees. Apple charges $99/year for a developer account; Google charges a one-time $25. On revenue, Apple takes 15% under €1M/year, 30% above; Google's standard commission drops to 20% on in-app purchases (10% on subscriptions) from June 30, 2026 in the US, UK, and EEA. If you sell through the app, the store is a permanent revenue-share partner — price it in.
Backend and hosting. A live app's servers, database, and APIs. Realistically €50–€500/month for a growing app on AWS, GCP, or a managed platform — scaling with users. A viral hit can push this into thousands a month, which is a good problem, but a real one.
Third-party API and SaaS subscriptions. Every integration that charged you to build also charges you to run: maps, messaging, analytics, email, push, error tracking, payment processing fees. Easily €100–€1,000+/month combined for a medium app, and it only grows with usage.
Updates for OS changes. New iOS and Android versions ship yearly and routinely break things. An app that fails on the latest OS gets uninstalled and review-bombed. This lives inside the maintenance budget, but it's the line that surprises people most.
Add it up: a €60,000 medium app realistically costs €12,000–€25,000/year to run between maintenance, hosting, APIs, and store fees — plus a revenue cut if you sell through the app. If a quote ignores all of this, it isn't a full quote. The build is the deposit; the run is the mortgage.
Here's the part most agencies won't say plainly, because it lowers their first invoice: don't build everything at once. The most expensive mistake in app development isn't hiring the wrong team — it's building the whole product before you know whether anyone wants it.
An MVP is not a cheap, broken version of your app. It's the smallest complete thing that proves the core idea works — the two or three features that matter, built properly, shipped to real users, on one platform. You learn more from a focused app in a thousand hands than from a feature-complete app in your imagination.
This ties directly to how we build. We add backend infrastructure when you actually need it, not by default. Realtime sync, multi-provider payments, an admin panel, both platforms — every one is real money, and every one is wasted if the core idea doesn't land. Ship the thin version, watch what people do, then invest in the parts that earned it.
The math is plain. A €120,000 full build that flops cost €120,000 and taught you one thing: it flopped. A €30,000 MVP that flops cost €30,000 and taught you the same thing with €90,000 still in the bank — for the next idea, or for building the right features once users showed you which ones matter. The MVP isn't the cheap option. It's the smart option, and it usually makes the eventual full product better and cheaper too.
The rule we live by: build the smallest thing that can prove or kill the idea. Then let real usage — not a feature wishlist — decide where the next euro goes.
Forget industry averages. Budget from your own answers to four questions:
1. What is the one thing this app must prove? If you can't name it in a sentence, you're not ready to build — you're ready to scope. The clearer the core, the smaller and cheaper the first version.
2. Does it really need a custom backend on day one? Often the answer is no. A thin backend, or even a manual process behind a polished front end, gets you to real users for a fraction of the cost. Add the expensive infrastructure when usage demands it.
3. One platform or two, native or cross-platform? Start with the platform your users are actually on, build cross-platform unless you have a specific reason not to, and you've already saved 35–50% before writing a line of feature code.
4. Can you fund the run, not just the build? Budget the build, then add 15–20% per year for maintenance and the monthly hosting/API stack before you commit. An app you can build but can't afford to run is a liability with an icon.
A sane process is: pick the smallest scope that proves the idea, pick the build method your budget and risk tolerance support, choose cross-platform unless you have a real reason not to, and add the recurring costs before you sign — not after. See our services for how we scope this in practice.
The studios worth hiring won't quote you blind. Discovery comes before the number — a real scope is the contract, and the price follows from it. Anyone who fires back a figure before understanding what your app needs to do is guessing, and you'll pay for the guess later.
Want a real number for your specific app instead of a range? Start a conversation — tell us what it needs to do, and we'll reply within 48 hours with a scoped quote before any work begins.
We take on a small number of projects each quarter. Tell us what you're building.
Astro's islands architecture produces the fastest marketing sites we've measured. Here's what the speed gains actually look like, when the learning curve is worth it, and when Next.js is still the better call.
The common objection is 'it's just a brochure site.' Here's why TypeScript on a marketing site pays off — through CMS integration, team handoffs, and the maintenance reality 12 months after launch.