How much does it cost to build an app in 2026?

A cross-platform mobile app that real people use costs $20,000 to $75,000 to build. A simple internal tool that replaces a spreadsheet or a paper form costs $6,000 to $25,000. A web application or customer portal costs $12,000 to $60,000. Those are our published ranges, and they hold up across most of the app work we quote from our Edina office.

If you want the honest version of the rest of this article in one line: the number depends almost entirely on how much you insist on building before you have a single user. Everything below is about narrowing that range for your specific project, and about the cheaper answer most people should take first.

What does an app actually cost in 2026?

Here is the starting grid. These are the ranges we publish on our pricing page, and the timelines are calendar time from kickoff to launch, not developer hours.

What you are building	Typical cost	Timeline	Who it is for
Interactive prototype only	$3,000 – $8,000	2 – 4 weeks	Testing an idea before committing budget
Internal tool or automation	$6,000 – $25,000	4 – 10 weeks	Replacing a spreadsheet, form, or manual step
Web application or portal	$12,000 – $60,000	8 – 24 weeks	Client portals, booking systems, member areas, dashboards
Mobile app, iOS + Android	$20,000 – $75,000	12 – 28 weeks	A first real version published to both stores
Mobile app with heavy backend	$75,000+	6 – 12 months	Marketplaces, fintech, real-time logistics

The low end of each range is a genuinely narrow product: one user type, one core job, no exotic integrations. The high end is several user roles, an admin system, payments, and half a dozen outside systems to talk to.

And nobody can give you a real number from a paragraph description. When someone quotes an app in a first phone call, they are either quoting a template or planning to make the difference back in change orders. A serious quote follows a scoping conversation where somebody asks uncomfortable questions about who logs in and what happens when they do.

Where does the money actually go?

App budgets get treated like a single lump of “development.” They are not. On a mid-range project, writing app code is under half the invoice. Here is the rough allocation on a $45,000 cross-platform build.

Phase	Share of budget	On a $45,000 project	What it buys
Discovery and product definition	8 – 12%	$3,600 – $5,400	User flows, feature cut list, technical architecture, the “no” decisions
UX design and prototype	10 – 15%	$4,500 – $6,750	Wireframes and a clickable prototype you can hand to real users
UI design	8 – 12%	$3,600 – $5,400	Visual design system, every screen state, icons, empty and error states
Frontend app build	25 – 30%	$11,250 – $13,500	The actual iOS and Android app from one codebase
Backend, API, database	20 – 30%	$9,000 – $13,500	Accounts, data, business logic, admin tools, security
Third-party integrations	5 – 15%	$2,250 – $6,750	Payments, CRM, EHR, calendars, SMS, maps, analytics
QA and device testing	8 – 12%	$3,600 – $5,400	Real devices, real OS versions, edge cases, accessibility
Store submission and launch	2 – 5%	$900 – $2,250	Store listings, screenshots, privacy disclosures, review responses

The two line items people try to cut are the first two, and those are the two that decide whether the other 75% gets spent on the right thing. Cutting discovery to save $4,000 is how a $45,000 app turns into a $70,000 app.

Backend cost surprises people. If your app needs users to log in and see their own data, you are not building one thing, you are building two. There is a server, a database, an admin interface, and an authentication system behind the pretty screens. An app with no backend, like a calculator or a reference tool, can be genuinely cheap. An app where two users see different content cannot.

How much does each type of app cost?

Category matters more than screen count. A booking app with fifteen screens costs less than a marketplace with ten, because the marketplace has two sides, money moving between strangers, and disputes.

App type	Realistic first version	What drives the cost
Internal tool or staff app	$6,000 – $25,000	One user type, private distribution, no store review
Content or community app	$15,000 – $40,000	Publishing workflow, notifications, moderation
Booking and scheduling	$25,000 – $60,000	Calendar logic, availability rules, reminders, cancellations, payments
Client or patient portal	$20,000 – $60,000	Roles and permissions, document handling, audit trails, compliance
On-demand or dispatch	$50,000 – $120,000	Live location, driver and customer apps, dispatch console, maps billing
Two-sided marketplace	$60,000 – $150,000	Two apps in one, payouts, ratings, trust and safety, disputes
Fintech or anything holding money	$80,000 – $250,000+	Identity verification, ledgers, audits, licensing, security review

Those top rows are not us posturing. If your idea is genuinely a marketplace or a payments product, the build is the smaller of your problems, and you should be talking to people who have shipped in that category specifically, not to a general web studio. We would tell you the same thing on a call.

What drives the price up the fastest?

In rough order of damage to your budget:

A custom backend instead of an off-the-shelf one. Using a managed platform for authentication, database, and file storage can save $8,000 to $20,000 on a first version. It is the single largest lever available to a small budget.

Real-time features. Live chat, live tracking, collaborative editing, anything where two screens must agree within a second. Real-time roughly doubles the engineering effort of the feature it touches, and it makes testing much harder.

Payments. Taking a card in-app is a few thousand dollars of work. Splitting money between multiple parties, handling refunds, subscriptions, failed renewals, and tax is a genuine subsystem. And on iOS, if you are selling digital goods consumed inside the app, Apple’s in-app purchase rules and commission apply, which changes your unit economics before it changes your build cost.

Offline mode. “It should work without signal” sounds like a small ask. It means your app now has two copies of the truth and needs rules for what happens when they disagree. Budget 20% to 40% on top of any feature that has to work offline.

Third-party integrations. Each one is a variable cost, not a fixed one. A well-documented modern API might be a day. An EHR or a state reporting portal with a PDF spec from 2011 might be three weeks. Get the API documentation in front of your developer before you sign anything.

Compliance. HIPAA, FERPA for schools, 245D documentation requirements for Minnesota waiver providers, PCI for cards, WCAG for accessibility. Compliance rarely adds a feature. It adds constraints on hosting, logging, access control, and process, and it typically adds 15% to 30% to a project’s cost. It is also not optional, so find out early. Our accessibility guide covers the WCAG side of that for smaller organizations.

Number of user roles. This is the sneaky one. Every distinct role — client, staff, supervisor, admin — is close to its own mini-application with its own screens, permissions, and test cases. Going from one role to three does not add 20%. It often adds 60%.

Native or cross-platform: what is the real cost difference?

For almost every organization reading this, cross-platform is the right call, and it is the cheaper call.

Approach	Cost vs. baseline	When it makes sense
Cross-platform (React Native, Flutter)	Baseline	90% of business apps: portals, booking, content, internal tools
One native platform only (iOS or Android)	15 – 25% less than cross-platform	You are certain your entire audience is on one platform
Two separate native apps	60 – 100% more	Heavy graphics, complex camera or sensor work, deep OS integration

Building two native apps means two codebases, two developers with different skill sets, two sets of bugs, and two maintenance bills every year for as long as the app lives. Shipping both stores from a single cross-platform codebase typically adds only 10% to 20% over building for one platform.

The honest caveat: cross-platform frameworks are worse when your app is essentially a camera, a game, an audio processor, or something leaning hard on brand-new OS features. If that is your product, pay for native and do not let anyone talk you out of it. We go deeper on this trade-off in native vs cross-platform.

One more thing that is not a technology decision at all: build for the platform your users actually have. Community organizations serving immigrant and lower-income households in the Twin Cities frequently skew heavily toward Android, and toward older devices. Building an iOS-first app for that audience is an expensive way to reach the wrong half of your users. Check the analytics on your existing website before you decide.

Offshore, onshore, or hybrid — what do you really pay?

Rates vary enormously, and so does what you get for them. Here is a fair picture, blended across a full team rather than a single developer.

Model	Blended rate	$45,000 buys you	Real risks
Offshore agency	$25 – $50/hr	900 – 1,800 hours	Time zones, specification gaps, rework, weak product judgment, handover problems
Hybrid (local lead + offshore build)	$60 – $95/hr	470 – 750 hours	Only works if the local lead is genuinely accountable for the code
US small studio	$95 – $175/hr	260 – 470 hours	Smaller team, less capacity for very large scope
US large agency	$175 – $300/hr	150 – 260 hours	Overhead, junior staff on your project, slow change requests

Our own hourly range is $95 to $150, which puts us in the small studio row.

Offshore is not a scam and it is not automatically a bad choice. Plenty of excellent software gets built in Poland, Argentina, Vietnam, and India. What offshore reliably does not give you at the low end is product judgment, meaning someone who pushes back and says “that feature is not worth $9,000, cut it.” You get exactly what you specified, which is fine if your specification is excellent and expensive if it is not.

The most common failure I see is not bad code. It is a fully built app that nobody thought carefully about, delivered on time, that does the wrong thing. That failure costs the whole budget, not part of it. If you go offshore, insist on three things: you own the repository from day one, you get a working build every two weeks, and there is one named person awake during at least four of your working hours. Our post on how to build an app without wasting $50k covers the contract side.

You probably do not need an app

This is the section that costs us money, so I will be direct about it.

Most organizations that come to us asking for a mobile app should build a fast mobile website or a progressive web app instead, and they should do it for $3,500 to $9,000 rather than $40,000. We say this on discovery calls several times a year, and we have talked people out of app projects we could have billed.

Here is the test. An app earns its cost when it clears at least two of these:

• People use it more than weekly, not once a year
• It genuinely needs the device: camera, GPS in the background, offline data, biometrics
• Push notifications drive real behavior, and email or SMS would not
• You need to live on the home screen because you are competing for habit
• You are shipping to a captive audience who will be told to install it, like staff or enrolled families

If you cannot clear two, you are buying a $40,000 icon. The most common version of this: an organization wants an app so people can find their hours, read their news, fill out a form, and donate or pay. Every one of those jobs is done better by a web page, because nobody downloads an app to check your hours once.

The cheaper path is not a downgrade. A modern mobile web app can be installed to the home screen, send push notifications on both major platforms, work offline for cached content, and load in under two seconds. It has no app store review, no $99 annual fee, no forced SDK upgrades, and no friction at the install step. You update it at 4pm and every user has the new version at 4:01, which is not true of native apps.

We would rather sell an organization a $6,500 web app or an excellent mobile-first site than a $45,000 native app that gets 40 installs. If the web version gets real usage in six months, the native app becomes an easy decision backed by data instead of a guess. If it does not, you saved $38,500 and learned the same lesson.

For internal tools there is an even cheaper honest answer: check whether a no-code platform solves it before you pay anyone to write code. Airtable, Glide, Softr, Jotform, and the automation features already inside your existing systems handle a real share of the “we need a simple tool” requests we get. If a $50/month subscription and a week of your own configuration replaces the spreadsheet, do that. Come back to us when you hit the ceiling: permissions the platform cannot express, an integration it does not have, or per-user pricing that has quietly grown past what custom would have cost. Plenty of organizations do hit it. Just do not start there.

The comparison between these three options in more depth is in web app vs website vs mobile app and website or app.

What does an app cost after it launches?

The build price is not the price. Apps decay in a way websites do not, because Apple and Google change the ground under them every year.

Ongoing cost	Typical amount	Notes
Apple Developer Program	$99 per year	Required to publish or keep an app on the App Store
Google Play developer account	$25, one time	Registration fee, not annual
Backend hosting and database	$25 – $500 per month	Scales with users and data; small apps sit at the low end
Push notifications, SMS, email	$0 – $200 per month	Push is usually free; SMS is metered and adds up quickly
Maps, storage, other metered APIs	$0 – $300 per month	Maps in particular can spike with usage
Maintenance and updates	15 – 25% of build cost per year	The line people forget

That maintenance figure is the one that matters. On a $40,000 app, that is $6,000 to $10,000 a year, and it is not padding. iOS and Android ship major versions annually. Apple periodically raises the minimum SDK version required for updates. Frameworks and dependencies get security patches. Third-party APIs get deprecated on somebody else’s schedule.

An app that gets no maintenance for two years does not gradually degrade. It hits a wall: it stops building, it gets pulled from a store for an outdated SDK, or a dependency has a security hole you cannot patch without upgrading everything at once. At that point the “cheap” path costs more than the original build.

Also plan for store review. Both stores review submissions, most come back within a day or two, and rejections are normal on a first submission — usually over privacy disclosures, account deletion, sign-in requirements, or a metadata detail. Budget a week of calendar time for the back and forth, and do not schedule a launch event before you have approval in hand.

If you sell digital goods or subscriptions inside the app, factor the store commission into your model. Both stores take a cut of in-app digital purchases, with reduced rates for smaller developers, and the rules around linking to outside payment have been shifting through recent US litigation. Physical goods and real-world services booked through an app are treated differently. Get this right at the business model stage, because it can change whether the app makes sense at all.

Are AI coding tools making apps cheaper in 2026?

Somewhat, and less than the marketing suggests.

What genuinely got cheaper: boilerplate, first drafts of screens, test scaffolding, data migrations, documentation, and the tedious work of wiring an API into an interface. A competent developer with good AI tooling moves noticeably faster through the mechanical middle of a project. We use these tools daily and they are real.

What did not get cheaper, because it was never typing:

• Deciding what to build and what to refuse to build
• Understanding a client’s actual workflow well enough to model it
• Integrating with the badly documented system your industry runs on
• Testing on real devices with real users who do unexpected things
• Security review, accessibility, compliance, and store submission
• Being accountable when something breaks at 9pm

That is most of the cost. So the honest 2026 number is that AI tooling has compressed build timelines and taken maybe 10% to 20% off the middle of a well-scoped project. It has not turned a $45,000 app into a $9,000 app, and anyone quoting as if it has is either skipping your discovery work or planning to hand you generated code nobody on their team fully understands.

There is a second effect worth naming. AI tools make it easy to produce a lot of code quickly, which means the projects that fail now fail bigger and faster. The discipline of cutting scope matters more than it used to, not less. That is as true of mobile app work as it is of anything else we build.

How do you get a real number for your project?

Write down three things before you talk to anyone. It will take an hour and it will save you thousands.

1. The one job. In one sentence, what does the app do that nothing you own does today? If you need three sentences, you have three projects.
2. Who logs in. List every distinct type of user and what each one can see and do. This single list drives more of the price than any other input.
3. What it must connect to. Name the systems, and find out whether they have an API. If the answer is “we email a spreadsheet,” say so — that is a fine answer and it changes the plan.

Then get quotes on the prototype first, as a separate paid engagement, before committing to the full build. You should be able to buy a clickable prototype for $3,000 to $8,000 that you can put in front of ten real users. Every app project we have run that started with a prototype ended with a shorter feature list than it started with. That is the point.

We scope software in phases with a fixed price per phase, and you can stop after any of them. That is not a courtesy, it is how you avoid a $60,000 commitment made on week one information. The same logic applies to websites, which we cover in how much a website costs in Minnesota.

Getting a straight answer

If you have an app idea and you want an honest read on whether it should be an app at all, get in touch or call the office at (612) 293-0636. We do mobile builds and web applications, and we tell people to build the cheaper thing first often enough that it is basically our default recommendation. A 30-minute conversation will tell you which range on that first table you are actually in, and whether you should be spending the money this year.