Key takeaways
- Price = hours × rate. Scope sets the hours, so compare quotes on scope, not on the headline number.
- A one-workflow web MVP is roughly 80–200 hours of senior engineering. Multiple roles, payments, native apps or compliance can double or triple that.
- Running a Cloudflare-first MVP usually costs tens of dollars a month in infrastructure, not hundreds. Paid third-party services and maintenance time cost more.
- The cheapest MVP is the one you don't have to rebuild. Buy auth, payments and email; build only the workflow that makes your product different.
Founders usually ask "how much does an MVP cost?" hoping for one number. The honest answer is a formula, and once you see the formula you can check any quote you get, including mine.
MVP cost = engineering hours × hourly rate (+ design, + third-party services). The rate varies by who you hire and where. The hours vary by scope, and scope is the part you control.
The short answer
For a web MVP with one core workflow, such as a booking flow, a client portal or a simple marketplace transaction, expect roughly 80 to 200 hours of senior engineering work. What that costs depends on how you buy it:
| How you buy it | Typical rate | 80 hours | 200 hours | Who carries overrun risk |
|---|---|---|---|---|
| Offshore freelancer | $25–$50/h | $2,000–$4,000 | $5,000–$10,000 | You |
| Senior freelancer (US/EU clients) | $75–$150/h | $6,000–$12,000 | $15,000–$30,000 | You |
| Small agency team | $100–$200/h blended | $8,000–$16,000 | $20,000–$40,000 | Usually you, via change requests |
| Fixed-price build | Flat fee | Fixed | Fixed | The builder |
Those rates are illustrative ranges, not survey data, so plug in the rates you're actually quoted. The point of the table is the multiplication: if two quotes differ by 3×, the difference is almost always hours (scope and team size), not the rate.
Where the hours actually go
Here is a realistic split for a single-workflow web MVP built by one senior engineer. Your numbers will differ, but the shape rarely does: the core workflow is the biggest line, and the "boring" parts add up faster than people expect.
| Line item | Hours (one workflow) | What it covers |
|---|---|---|
| Scoping and technical plan | 6–12 | User flow, data model, integrations, what is out of scope |
| Project foundation | 8–16 | Repo, environments, CI, deploy pipeline, error tracking |
| Auth and accounts | 6–16 | Sign-up, login, sessions, password reset, one role |
| Core workflow (frontend + API + data) | 30–90 | The thing users came for, end to end |
| Integrations | 8–30 | Payments, email, calendar, a third-party API |
| Basic admin and support tools | 4–16 | Usually a protected page or a database console, not a full admin panel |
| Testing, fixes and launch | 10–24 | Production checks, edge cases, deploy, monitoring |
| Handover | 3–8 | Docs, account transfer, walkthrough |
Two things stand out. First, the core workflow is a range of 30 to 90 hours on its own, which is why "how complex is the main flow?" matters more than anything else. Second, accounts, deployment and launch work are fixed costs: you pay them whether your workflow is tiny or huge.
What makes an MVP budget balloon
Most budgets double for the same handful of reasons. If your idea includes any of these, plan for them explicitly or cut them from version one.
- More than one user role. Buyers and sellers, patients and clinicians, admins and members. Each role brings its own screens, permissions and edge cases. Two roles is often 1.5× the work of one.
- Payments with real billing logic. A one-off Stripe Checkout is a few hours. Subscriptions, trials, proration, usage billing, refunds and invoices are a project of their own.
- Native mobile apps. iOS and Android add app store review, device testing and a second (or third) client. A responsive web app tests the same idea for a fraction of the cost.
- Real-time features. Live chat, collaborative editing and live dashboards need persistent connections, presence and conflict handling.
- Compliance. HIPAA, SOC 2 or strict GDPR requirements change how data is stored, logged and accessed. I've built HIPAA-ready backend foundations for aa.health; it's very doable, but it is not free.
- Migrating existing data. Importing a spreadsheet is easy. Reconciling records from an old system with different rules is not.
- A custom design system. Bespoke UI for every component looks great and costs weeks. Start from a solid component library and customise the parts users actually see first.
- Integrations with poor APIs. A well-documented API takes hours. An undocumented partner API with rate limits and flaky webhooks can take days.
Running costs after launch
The build is a one-off. These costs keep coming every month, and founders often forget to budget for them.
Infrastructure
On a Cloudflare-first stack, early-stage infrastructure is cheap. Cloudflare's Workers Paid plan starts at $5 a month and covers a lot of traffic before usage charges apply, and R2 object storage has no egress fees, which matters as soon as you serve files. A small database on D1 or a managed Postgres starter plan adds little at MVP volume. Check each provider's current pricing page, because plans change, but expect tens of dollars a month, not hundreds, until you have real traction.
For context, I've run Cloudflare-first production backends at much larger scale: Whydonate handles 4M+ API requests a day on Workers, Queues, R2, PlanetScale and Typesense. The architecture that keeps an MVP cheap is the same one that scales, if each piece has a clear job.
Third-party services
These usually cost more than hosting:
- Transactional email (sign-up, password reset, receipts)
- An auth provider, if you don't run auth yourself
- Error tracking and uptime monitoring
- Payment processing fees (a percentage of revenue, not a subscription)
- Domain, and any paid APIs your workflow depends on
Maintenance time
Dependencies need updates, a payment provider changes an API, a user finds an edge case. Budget a few hours a month for a live MVP, more in the weeks after launch when real users show you what you missed.
How to spend less without shipping something worse
The goal isn't the cheapest build. It's spending money only on what tests your idea.
- Build one workflow. Pick the single path a user takes to get the core value, and build that end to end. Everything else waits. I wrote a whole guide on how to scope an MVP you can launch in two weeks.
- Buy the commodity parts. Auth, payments, email and file storage are solved problems. Use a provider; build only what makes your product different.
- Skip the admin panel. For the first hundred users, a protected page or your database console is enough. Build admin tools when you know which ones you need.
- Start on the web. A responsive web app tests demand on every device. Go native when you have a reason.
- Choose boring, cheap infrastructure. A serverless stack (Cloudflare Workers, or AWS Lambda where it fits) costs almost nothing at low traffic and doesn't need a rewrite when traffic grows.
- Fix the scope before the price. A fixed scope with a fixed price moves overrun risk to the builder, which is where it belongs, since they control the estimate.
What a $4,500 fixed-price MVP includes
My own offer is built on that formula. I build a focused web MVP in 14 days for a flat $4,500: one core workflow, a Next.js frontend, APIs, auth, a database and the core integrations, deployed to production (Cloudflare-first where it fits), with the source code, accounts and handover notes belonging to you.
What it deliberately leaves out: native mobile apps, multi-role marketplaces with complex permissions, heavy compliance work and anything that's really two products. If your idea needs those, I'll say so on the first call and we'll either cut it to a launch slice or scope a longer build.
A five-minute budget worksheet
Before you request quotes, write down answers to these. They're also the questions any good builder will ask you.
- What is the one workflow a user must complete for your product to be useful?
- How many user roles does that workflow need? (Aim for one.)
- Which integrations are essential on day one, and which can wait?
- Do you need payments at launch, and are they one-off or recurring?
- Is there regulated data (health, finance, children's data) involved?
- Web only, or are native apps truly required for version one?
- What's the launch date, and what happens if you miss it?
Bring those answers to a quote and you'll get comparable numbers. Bring them to me and I'll tell you on a 20-minute call whether the idea fits a 14-day build.
Frequently asked questions
Is $5,000 enough to build an MVP?
For a focused web MVP with one core workflow, yes, if the scope is fixed and the builder is senior enough not to waste hours. It is not enough for a multi-role marketplace, native apps or a product with heavy compliance requirements. The scope decides, not the number.
How long does it take to build an MVP?
A single-workflow web MVP can launch in two to six weeks. Two weeks needs a locked scope and one experienced engineer; more roles, integrations or platforms push it towards two or three months.
Should I use no-code instead?
No-code tools are great for validating demand with a landing page or a manual concierge version. Once you need custom logic, reliable integrations or data you'll want to keep, a small coded MVP is often cheaper over the first year than fighting a platform's limits.
What does it cost to maintain an MVP?
Expect infrastructure in the tens of dollars a month on a serverless stack, third-party services on top, and a few hours of engineering a month to keep dependencies and integrations healthy. Plan for more in the first month after launch.
Do I own the code when I hire a freelancer?
You should. Make code ownership, repository access and deployment accounts part of the agreement before work starts. In my builds, the source code and deployment accounts belong to you.