MVP development in London, fixed price, live in four to six weeks.

An MVP is the one product where the buyer knows the word before they know the price, and where the market has settled on a shape: one core workflow a paying user will actually use, sign-up and accounts, a way to take money, an admin view for the founder, and enough instrumentation to learn from the first hundred users. Everything else is a second phase. The failure mode is the opposite: six months and a feature list, launched to nobody.

What is not in an MVP matters as much. We will not build a native iOS and Android app into a first release unless the product cannot exist without it; a responsive web app reaches the same users for a third of the cost and can be wrapped later. We will not build a second user type, a marketplace, or an integration marketplace into a first release either. Those are the things founders are quoted £40,000 for, and they are the reason most MVPs miss the money they were meant to raise.

We price a startup MVP at £10,000 to £20,000, fixed against a written scope, and take four to six weeks to a launch that real users can pay for. That sits where the London market sits. Magora, a London studio since 2010, publishes "Startup MVP: from £10,000" and "MVP in 6 to 8 weeks". Coderacle, also London, publishes a £15,000 fixed-price production SaaS MVP with a median of eight weeks to launch. Wider surveys of London software firms put an MVP higher again: Red Eagle Tech publishes £10,000 to £35,000 and Square Root Solutions £25,000 to £80,000.

The difference is not the number. It is that we have launched two products of our own and run them every day, so the architecture your MVP is built on is the one we would choose for ourselves, with cost instrumentation and billing metering in from the first commit rather than retrofitted at scale.

Competitor figures read on their own pages on 14 September 2026; prices exclude VAT where stated. All are credible choices; we list them because a founder should compare.

Week one: the scope and the acceptance criteria in writing, the one workflow agreed, and working software on real data before the week is out. Weeks two to four: the core workflow, sign-up and accounts, and payments, with a demo every week so nothing arrives as a surprise. Week five: the admin view, analytics and error tracking, then testing with real users, including the awkward cases. Week six: launch, handover of the repository, accounts and runbook, and a plan for what to learn from the first users.

Most delays on a first product are decisions waiting on somebody, not engineering. That is why one named contact on each side, a written scope and a weekly demo matter more than the size of the team.

Add a model when it does the job the product exists for: reading a document, scoring a CV, drafting a reply, answering a question from your own data. Do not add one because investors like the word. An AI feature has a running cost per user that a conventional feature does not, and an MVP priced per seat can lose money on its heaviest users without anybody noticing.

So every AI feature we build into an MVP is attributed to a customer and a feature from the first release, with a spending ceiling that has been watched firing before launch. The small model does most of the work; in our own products, moving one high-volume feature from the frontier model to the small one made it several times cheaper and faster with no drop in success.

GoRizzume is an AI career platform we built and run: CV scoring against a specific advert, spoken interview practice and a UK licensed sponsor register of more than 120,000 sponsors, refreshed every day. It started as a web product, then added iOS and Android apps and a browser extension, and it is licensed to institutions by seat as well as sold to individuals.

Onrolo is a recruitment operations platform we built and run, on web, iOS and Android, in daily use across a national recruitment network. It removes 90 or more hours of manual CV screening per office every month, and eight in ten applicants reach a booked appointment.

Both carry usage metering wired into billing and every model call attributed to a feature and a customer. That is the architecture your MVP starts on.

The repository, the hosting and the provider accounts are in your name from day one, not on final payment. At launch you receive documentation, credentials, a runbook and a recorded walkthrough, so another developer could pick the product up without speaking to us.

That is also what technical due diligence checks: who owns the code and the accounts, whether the history is in your repository, what the product costs to run per user, and whether anyone other than the original builder could maintain it. We build so that each of those questions has a short answer.

There are four common routes to an MVP in London. The right one depends on how much you can specify, how quickly you need to launch, and how much you need to own at the end.