How Much Does It Cost to Build an MVP?
A focused web or SaaS MVP built by a small senior team costs roughly $1,000 to $2,000 and takes three to six weeks. Mobile apps land in a similar range. The price is driven by the number of distinct user flows and third-party integrations, not by how many screens the product has, which is why cutting features often saves far less money than cutting one user role.
The gap between the cheapest and most expensive quote for the same MVP is routinely ten to one. That spread is not about quality. It is about what each quote assumes is in scope, and almost nobody writes that part down.
Here are real numbers and what actually moves them.
What an MVP costs
This is what we publish, so you have something concrete to calibrate against:
| Build | Price | Timeline |
|---|---|---|
| Something smaller, one clear job | $150 to $500 | 2 days to 1 week |
| AI chatbot or concierge | $500 to $1,250 | 2 to 3 weeks |
| Mobile app | $750 to $1,500 | 3 to 6 weeks |
| Web app or SaaS MVP | $1,000 to $2,000 | 3 to 6 weeks |
| Ongoing changes and support | $250 to $750 per month | Monthly |
Every price is a range with a condition attached. A single number quoted before anyone has described how the product is used is a number that gets renegotiated later.
Traditional agencies commonly quote ten to twenty times these figures for the same scope. The difference is account management, a discovery phase that produces a document, design rounds, and the margin covering all three.
What drives the price
User roles, first
A role is a kind of person who sees something different. A customer, a shop owner, an admin. Each brings its own screens, permissions, test cases and failure modes. Adding a second role costs more than adding five screens to the first.
Most products that think they need three roles at launch need one. The other two can be you, working directly in the database, for the first month. At eleven customers that is the correct engineering decision, not a shortcut.
Integrations, second
Anything you do not control. Stripe, a calendar, WhatsApp, a CRM, a shipping provider. Each brings authentication, webhooks, retries, sandbox testing and failures that only appear in production. One is normal. Three is where timelines slip.
What costs less than people expect
- Screen count. Twenty screens across two roles is cheaper than eight across six roles.
- Visual polish. A real design system from day one is close to free.
- Framework choice. Any competent modern stack lands within a few percent of any other.
Should you build it in Lovable first?
Worth taking seriously. Lovable, Bolt and Replit get you a working interface in days for almost nothing, which is genuinely useful for testing whether anyone wants the thing.
The cost arrives later. These tools produce working React and Supabase code but do not reliably write restrictive database permissions, keep API keys off the client, or validate on the server. That becomes a hardening pass before real customers use it, currently $150 to $400 for a review and up to $1,500 with fixes.
So the honest comparison is not $0 against $1,500. It is roughly $500 plus a $1,000 hardening pass against $1,500 built properly from the start. At those numbers, building it properly costs about the same as patching it later, so the question is really whether you still need to prove demand before spending anything.
The failure mode is skipping the hardening pass because the app appears to work. The problems are not visible from using it. Your Lovable or Claude-built app is not production ready yet is the checklist.
How to scope so the budget holds
Write the scope as flows, not features. A feature list is a wish. A flow is testable: "a customer books a class and pays for it" either works end to end or it does not.
Then cut, in order:
- Roles. Down to one if you can.
- Integrations, to one. Keep the one without which the product does not function. Everything else is a manual step you do yourself while the numbers are small.
- Settings pages. Hard-code the choice. Ship. Find out whether anyone asks.
Do not cut: anything touching money or personal data being correct, and the shape of your data model. Both are cheap now and expensive in six months.
What a real quote contains
- Scope as flows, plus an explicit out-of-scope list. The out-of-scope list is the more useful half.
- Named milestones you can open in a browser. Not percentages.
- Ownership in writing. Repository, cloud accounts, domain, database and API keys under your organisation, with the builder granted access. Recovering access from a previous developer takes weeks and blocks everything else.
- A post-launch bug policy. What counts as a bug rather than a change, and for how long it is fixed free.
- A rate for out-of-scope work. Agree the price of changing scope before you want to.
Hiring for it
What to search for depends on what you are building. "MVP development company" and "software development agency for startups" return the widest range of price and quality. More specific terms get you closer to someone who has done your exact thing: "Next.js developer", "React Native developer", "Supabase developer", "FastAPI developer", "AI integration agency".
How to filter quickly. Ask for two products you can open in a browser today, and ask who owns the code when the engagement ends. Those two questions remove most of the field. A studio that cannot show live work, or that wants to keep the repository, is not the one.
Hourly or fixed price. Fixed price for a scoped MVP, because the risk of scope being wrong should sit with whoever wrote the scope. Hourly for ongoing changes after launch, where the work genuinely cannot be specified in advance.
Budget for the month after launch
An MVP nobody can deploy, monitor or change is not finished. Before launch you want a deploy anyone on the team can run, error tracking that notifies a person, and a database backup you have restored once and actually looked at.
A day of work. Skipping it is the most common reason a product that launched fine is stuck three months later.
Related
Full-stack development and app development cover how we scope and run these. The class booking platform and the QR cafe ordering system were scoped exactly this way: one primary flow, one integration, everything else deferred.
If the plan involves a chatbot over your own content, what a RAG chatbot really costs prices that separately.
Common questions
- How much does it cost to build an MVP?
- A focused web or SaaS MVP from a small senior team usually costs $1,000 to $2,000, with mobile apps in a similar range and smaller single-purpose builds starting around $150. The lower end covers one user flow and one integration. The upper end covers several user roles, payments and an admin surface.
- How long does it take to build an MVP?
- Three to six weeks for a focused scope. Under two weeks usually means a prototype rather than something you can charge for, and over three months usually means the scope stopped being minimum some time ago.
- What makes an MVP more expensive?
- User roles first, integrations second. Each new role brings its own screens, permissions and test cases, so it costs roughly as much as a whole feature area. Each integration with Stripe, a calendar, WhatsApp or a CRM brings authentication, webhooks, retries and sandbox testing. Screen count and visual polish cost much less than people expect.
- Is it cheaper to build an MVP with Lovable or Bolt first?
- Cheaper to start, not always cheaper to finish. AI builders get you a working interface in days for almost nothing, which is genuinely useful for testing demand. The cost arrives later, as a hardening pass to fix database permissions, exposed keys and missing server-side validation before real customers use it. Budget for that pass rather than being surprised by it.
- How do I choose an MVP development agency?
- Ask for two products you can open in a browser today, and ask who owns the code when the engagement ends. Those two questions remove most of the field, because a studio that cannot show live work or wants to keep the repository is not the one. Search broadly with terms like MVP development company or software development agency for startups, then narrow with the specific stack you need, such as Next.js developer or React Native developer.
- What should an MVP quote include?
- Scope written as user flows rather than a feature list, an explicit out-of-scope list, named milestones you can open in a browser, who owns the repository and cloud accounts, what counts as a bug versus a change after launch, and the rate for out-of-scope work. A single number with no scope attached is not a quote.
Want this looked at properly?
Tell us what you're working with. We'll give you a straight answer in a 30-minute call, free.