Pricing
How Much Does It Cost to Build an App in the UK? (2026)
What a UK small business really pays to build an app or web app in 2026: price bands, what drives them, the running costs nobody quotes, and how to spend less.

Search for what an app costs in the UK and most of the answers come from agencies pricing iPhone and Android apps for funded start-ups. That is rarely what a small business needs. More often it is a booking system, a customer portal, a members’ area or a tool that replaces a spreadsheet, and those are usually best built as a web app.
This guide gives you our actual price bands, explains where the money goes, lists the running costs with real list prices, and shows how to make the first version smaller and cheaper. Where we use examples, they are products we designed and built for our own businesses (TenancyDesk, Bookstorez, GoldStarred and Cocoon), not client invoices.
The short answer: what an app costs to build
These are the bands we quote within. They are for a first release that does its main job properly, not a demo.
What you are building | First release | Usually includes |
|---|---|---|
Internal tool or single automation | £4,000–£12,000 | One team and one workflow, sign-in, connected to one or two tools you already use |
Customer-facing web app or portal | £7,000–£25,000 | Customer accounts, a handful of core screens, email notifications, an admin view, payments if needed |
Subscription product or marketplace | £20,000–£50,000+ | Plans and billing, several kinds of user, files or documents, a back office, security hardening |
Native iPhone and Android apps | Quoted on top | Store listings, reviews and two platforms to maintain. An installable web app often removes the need |
If those numbers feel high, the cheapest legitimate route is usually not a cheaper developer but a smaller first version. If they feel low, it is because a web app avoids building the same thing twice for two phone platforms.
First, which kind of app do you actually need?
“App” covers three quite different things, and the choice moves the price more than anything else.
A native mobile app
Built for iPhone and Android and installed from the App Store and Google Play. You pay to be in the stores: Apple’s developer programme is 99 USD per membership year, and a Google Play developer account has a US$25 one-time registration fee. If you sell anything inside an iPhone app, Apple takes a commission: 15% under its Small Business Program for developers with up to 1 million USD in proceeds, and its standard, higher rate above that. Every update also goes through store review.
A web app
Software that runs in the browser on any device. There is one version to build and maintain, updates reach everyone the moment you publish them, and nobody has to approve your release. Most customer portals, booking systems, back offices and subscription products are web apps.
An installable web app (a PWA)
A progressive web app is a web app that can be added to a phone’s home screen and opened like any other app, with an offline page when there is no signal. Cocoon’s members’ Club works this way: it is a progressive web app with a service worker, so members open it from the home screen, and it never went through an app store.
Go native when the product genuinely depends on the phone: heavy use of Bluetooth or background location, or when being found in the app stores is how you will get customers. For most business software, a web app or PWA does the job for less.
What actually drives the price
Any honest quote is an estimate of days multiplied by a day rate, plus some contingency. So the useful question is not “what does an app cost?” but “what makes it take more days?”. These are the decisions that do.
Decision | Why it adds work | How to keep it smaller |
|---|---|---|
Accounts and sign-in | Security, resets, sessions, who can see what | Sign-in by email link, with no passwords to manage |
Kinds of user | Each role needs its own screens and permissions | Launch with one kind of user and add the next later |
Taking payments | Checkout, receipts, failed cards, refunds, VAT | Use a hosted checkout page rather than building your own |
Subscriptions | Trials, upgrades, cancellations, chasing failed payments | One plan at launch |
Connections to other tools | Every outside system has its own rules and failure modes | One integration first; a spreadsheet export for the rest |
Documents and PDFs | Layout, exact wording, versions, storage | One document type first |
AI features | Instructions, checking the output, cost limits | AI drafts, a person approves |
A back office | Someone has to run the business behind the app | Start with the database’s own dashboard |
Our own products show how quickly these add up. Bookstorez signs every user in by email link, so there are no passwords to store or reset. TenancyDesk tracks 21 legal duties for every property and tenancy, generates four official documents as PDF and Word, and lets landlords and tenants sign online, with each signed agreement carrying a SHA-256 fingerprint that anyone can check. Each of those features is its own block of days, and a product of that shape would sit in our top band if a client commissioned it.
The running costs nobody puts in the quote
Building the app is a one-off. Running it is not. Before you commit, add up what it will cost each month to keep it online, take money and send email. TenancyDesk, Bookstorez and Cocoon’s Club all run on the same shape of system: a Next.js app hosted on Vercel, a Supabase database, Stripe for payments and Resend for email. Here are the list prices for that kind of setup.

What you pay for | List price | Notes |
|---|---|---|
Hosting (Vercel) | Hobby $0/mo; Pro $20/mo | The free tier is for personal projects; a business app needs a paid plan |
Database and file storage (Supabase) | Free $0/mo; Pro $25/mo | Usage beyond the plan’s allowance is charged on top |
Card payments (Stripe, UK) | 1.5% + 20p standard UK cards | 2.8% + 20p premium UK cards, 2.5% + 20p EEA cards, 3.15% + 20p international cards, plus 2% if currency conversion is needed |
App stores (native apps only) | Apple 99 USD a year; Google Play US$25 once | Plus Apple’s commission on paid apps and in-app purchases |
Email, AI and domain | Usage-based | Small volumes are cheap; set a monthly cap on any AI usage |
Changes and support | Agreed with your developer | Usually a monthly retainer; the biggest running cost over time |
Card fees are easy to underestimate because they scale with your sales. On a £9 monthly subscription paid with a standard UK card, Stripe’s 1.5% + 20p comes to about 34p, nearly 4% of the price, because the fixed 20p weighs heavily on small payments. On a £90 payment the same fee is £1.55, under 2%.
How much does it cost to maintain an app?
The honest answer is that an app is never finished, so maintenance is less a fixed fee than a budget for change. Libraries need updating, a payment provider changes an API, the law changes, and once real people use the app you will want things done differently.
Our own products show the pace. Counting on 14 November 2026, TenancyDesk’s code had taken 301 commits since its first on 5 September 2026; Bookstorez had taken 1,029 since 20 September; and GoldStarred 455 since 24 October. Not every commit is a feature, but every one is a change somebody decided was worth making.
So decide the monthly budget for changes before you sign for the build, not after launch, and make sure the code, hosting and accounts are in your name so you can take that budget elsewhere if you ever need to. (We cover ownership properly in our guide to bespoke versus off-the-shelf software.)
Can I build my own app for free?
You can get surprisingly far. No-code builders and AI app builders will let you make a working prototype without paying a developer, and most have a free tier to start on. That is a good way to test whether anyone wants the thing before you spend real money.
The trade-offs show up later: paid plans that rise with your number of users, limits on what the logic can do or connect to, and an app that lives on someone else’s platform rather than as code you own. A sensible path is to prove the idea with a no-code tool, then build properly once you know exactly what it needs to do.
What should a first version leave out?
The most reliable way to spend less is to launch with one job done well. We build our own products in slices. Bookstorez launched with shops, product pages and browsing first, with the Buy button switched off until payments were the next piece. GoldStarred’s first slice went live on a preview address on 24 October 2026, with taking payments left for a later phase.
Things that can usually wait for version two:
- A second or third kind of user
- Native iPhone and Android apps (start with an installable web app)
- Every integration except the one the job depends on
- Custom reports and dashboards (export a spreadsheet instead)
- Multiple pricing plans, coupons and annual billing
- A polished back office (start with simple admin screens)
What do AI features add to the cost?
Two things: build time, to write the instructions and check what the model produces, and a running cost, because model providers charge for each use. The safe pattern is to have AI do the tedious part and a person confirm it. TenancyDesk reads the dates off an uploaded gas safety record or EPC with AI, and the landlord confirms them before anything is saved. Bookstorez’s ebook generator flags anything that was not in the seller’s own notes so they can check it. Put a monthly spending cap on any AI feature from day one.
How to get a quote you can trust
Whoever you talk to, give them this and ask for this:
- Describe the job, not the features. “Customers book and pay for a cleaning slot, and we see the week’s bookings” gets a better estimate than a list of screens.
- List the kinds of user and what each one must be able to do.
- Name what it must connect to: accounts software, CRM, calendar, payment provider.
- Split must-haves from later. Ask for the first phase to be priced on its own.
- Ask how the price was built: days per feature, the day rate and what is excluded.
- Ask what it will cost to run each month, and who owns the code, the hosting and the data.
Frequently asked questions
How much does it cost to build an app in the UK?
For a first release in 2026, Nexa quotes within these bands, excluding VAT: £4,000–£12,000 for an internal tool or single automation, £7,000–£25,000 for a customer-facing web app or portal, and £20,000–£50,000+ for a subscription product or marketplace. Native iPhone and Android apps are quoted on top. Other studios price differently, so ask any supplier how their quote was built.
Is a web app cheaper than a mobile app?
Usually, yes. A web app is one system that runs in any browser, while native apps mean building for iPhone and Android and paying to be in the stores (Apple’s developer programme is 99 USD a year; a Google Play account has a US$25 one-time fee). A progressive web app can still be installed on a phone’s home screen. Native apps are worth it when the product depends on phone hardware or app-store discovery.
How much does it cost to run an app per month?
For a small web app, list prices on 14 November 2026 were $20 a month for Vercel Pro hosting and $25 a month for a Supabase Pro database, plus card fees (Stripe charges 1.5% + 20p on standard UK cards), email, any AI usage and a budget for changes. The budget for changes is usually the largest running cost over time.
How much does it cost to maintain an app?
An app is never finished: libraries need updating, providers change their APIs, the law changes and users ask for improvements. Treat maintenance as a monthly budget for change agreed before the build starts, and make sure the code, hosting and accounts are in your own name.
Can I build an app for free?
You can build a working prototype with no-code or AI app builders, most of which have a free tier. That is a good way to test demand. The trade-offs come later: prices that rise with users, limits on logic and integrations, and an app that lives on someone else’s platform rather than code you own.
How long does it take to build an app?
A useful first release typically ships in 4–10 weeks, depending on scope. The fastest way to shorten it is to launch with one job done well and leave extra user types, integrations and native apps for later phases.
What does an MVP cost?
An MVP (minimum viable product) costs whatever the smallest useful version of the job costs, which for most small-business web apps sits at the lower end of the bands above. Leave out second user types, extra integrations, custom reports and multiple pricing plans until real users have shown they need them.
Tristan Glaves
Founder, Nexa
Founder of Nexa. Background in software engineering and product. Builds bespoke software and digital products for UK businesses.
2–4 specific automation ideas for your business · No account needed