In 2026, MVP development in India costs roughly 50,000 to 1.5 lakh for a landing page test, 1 to 3 lakh for a clickable prototype, 3 to 8 lakh for a working product with one real feature, and 5 to 25 lakh for a first version you can actually sell. Most founders are quoted the last row and shown the words "minimum viable".
That substitution is the single most expensive thing in this article. A 20 lakh quote for an MVP is usually a quote for version 1.0, and version 1.0 is what you build after the MVP has told you something. The ranges above are not four prices for the same thing. They are four different questions, and picking the cheapest one that answers yours is the whole skill.
This guide is about scope and cost. How long each takes is in how long it takes to build an MVP, and the platform-specific build figures are in our app and SaaS cost guides.
The four rungs, and what each one buys
| Rung | What it is | Cost in India | What it actually tests |
|---|---|---|---|
| Landing page and manual service | A page, a payment or waitlist link, and you doing the work by hand behind it | 50,000 to 1.5 lakh | Whether anyone wants it enough to act |
| Clickable prototype | Designed screens wired together, no backend, put in front of real users | 1 to 3 lakh | Whether people understand it and can complete the flow |
| One real feature, working | Accounts, the single core action, and the data behind it. Nothing else | 3 to 8 lakh | Whether it works in their hands, with their data, repeatedly |
| Sellable first version | Sign-up, payments, admin, the support tools, the second user type | 5 to 25 lakh | Whether they will pay, and whether you can run it |
Rungs one and two are not lesser MVPs. They are faster answers to earlier questions, and skipping them is why the fourth row so often gets built for a market that was never confirmed. Our own products taught us this the expensive way, and the honest account is in four products in, what we got wrong.
The market's published ranges agree with the table: Indian agencies quote 2 to 5 lakh for a basic MVP, 5 to 12 lakh for a standard one and 12 to 25 lakh and up for an advanced one, at 20 to 50 dollars an hour. The disagreement is never about the rate. It is about which rung the word covers.
Where the money goes: a worked estimate
Take the third rung, which is where most funded MVPs should start: a working product with accounts and one core feature, used by real customers. Say a booking tool for a services business, where the core action is making and confirming a booking.
| Work package | Person-days |
|---|---|
| Discovery: the one question the MVP must answer, and what result would change your mind | 4 |
| Design: the core flow, the states, a clickable prototype tested on five people | 8 |
| Accounts: sign-in, profiles, the minimum permissions | 5 |
| The core feature, end to end, including the unhappy paths | 14 |
| Data and admin: somewhere to see what is happening and fix it by hand | 5 |
| Payments, only if money must change hands to answer the question | 5 |
| Instrumentation: the three numbers that tell you whether it worked | 2 |
| Testing on real devices, and a closed launch to the first users | 4 |
| Total | 47 |
At the blended 16,000 rupees a person-day we use across these guides, that is about 7.5 lakh, inside the third row, and eight to ten weeks with two people. Add the fourth row's items, billing, a second user type, support tooling, and it roughly doubles, which is exactly the gap between the third and fourth rungs.
Two lines deserve attention. Instrumentation is two days and it is the point of the whole exercise: an MVP that ships without the three numbers that decide its fate is just a small product. And payments are optional at this rung, which founders resist, because taking money feels like the proof. It is, but a payment link beside a manual invoice proves it for a tenth of the cost.
The seven things that move an MVP quote
- The number of user types. Customers and providers is not one MVP, it is close to two: two onboardings, two sets of screens, two permission models. Launch one side manual, as most marketplaces actually did.
- Whether money moves. Payments mean gateways, declines, refunds and reconciliation. If the question you are testing is demand rather than willingness to pay a particular way, defer it.
- How much of the unhappy path you build. The feature working is a third of the work. What happens when it fails, times out, or gets bad input is the rest, and it is not optional once real users arrive.
- Platform. Web only is the cheapest honest answer. Two native apps is the most expensive. One codebase for both is in between, and the trade-offs are in Flutter against React Native against native.
- Integrations. Every outside system is a week or more, plus someone else's approval timeline. An MVP with four integrations is not an MVP.
- Design depth. A design system is a version 1.0 cost. An MVP needs a clear flow and states, not a token library.
- Decisions not yet made. Changing the core flow in week six costs more than everything in the discovery line. A week spent agreeing what would count as failure is the cheapest insurance available.
What an MVP should not include
Said plainly, because every one of these appears in quotes we are asked to review:
- An admin panel with roles and permissions. You can run it from the database for fifty users. Build the panel when a person who is not you needs it.
- A second language. Unless the test itself is about that market.
- A notification system. One email on one event is enough until someone asks for the second.
- Analytics dashboards. You need three numbers, not a dashboard. A spreadsheet pulled weekly is fine.
- A mobile app, if the web works. Installation is a barrier for something nobody has used yet.
- Scale. Nothing in an MVP should be built for a load you have not got. The rewrite, if you earn one, is the good outcome.
Every item on that list is a real feature that a real product eventually needs. The question is never whether, it is whether before the thing has proved itself.
Buy, assemble or build
| Route | Cost | Fits |
|---|---|---|
| No-code assembly | 20,000 to 1.5 lakh, plus subscriptions | Testing demand and flow, internal tools, anything with under a few hundred users and no unusual logic |
| No-code plus a custom piece | 2 to 6 lakh | The process is standard except for the one thing that is yours |
| Custom MVP | 3 to 8 lakh for one feature, 5 to 25 lakh for sellable | Your logic is the product, the data is sensitive, or the volume breaks the tool's pricing |
The honest sequence for most founders is the first row, then the third. Going straight to custom is right when the thing you are testing is the software itself; it is wrong when you are testing whether anyone wants the service the software would deliver.
We say this to people who are asking us to quote for a build, and it costs us work. It is still the right advice.
What the same MVP costs elsewhere
Founders in London and New York ask us this version of the question, so here are the comparable numbers for the third rung, a working product with accounts and one core feature.
| Where | Typical cost | Why |
|---|---|---|
| India | 3 to 8 lakh, about 3,500 to 9,500 dollars | Blended rates of 20 to 50 dollars an hour |
| Eastern Europe | 25,000 to 50,000 dollars | Rates of 50 to 90 dollars an hour |
| United Kingdom | 40,000 to 90,000 pounds | Agency day rates of 600 to 1,200 pounds |
| United States | 80,000 to 200,000 dollars | Agency rates of 150 to 250 dollars an hour |
The work is the same 47 person-days in each row. What differs is the rate, and the rate is mostly a function of where the people live. We set out the fully loaded comparison, including the costs of managing a team in another time zone, in our guide to hiring developers in India.
The honest caveat: the gap is real but it is not free. An offshore MVP needs the same founder attention as a local one and a bit more writing, because the corridor conversation that fixes a misunderstanding in thirty seconds does not happen. Teams that work well across time zones overlap deliberately for a few hours a day and put decisions in writing. That is a habit, not a cost, and it is the one thing that decides whether the arbitrage survives contact with the project.
The costs that are not in the quote
| Cost | Typical size | When it surprises you |
|---|---|---|
| The month after launch | 15 to 25 percent of the build | When the first real users find what testing did not |
| Running it | 3,000 to 25,000 a month for hosting, email, SMS and gateways | The first invoice |
| App store accounts and review | 10,000 a year, plus days of waiting | When the launch date meets a rejection |
| Your own time | A day a week, minimum, for the whole build | Always. An absent founder is the most expensive line in an MVP |
| The pivot | A few lakh, cheerfully spent | When the MVP works and tells you something you did not expect |
| Version 1.0 | The ranges in our app and SaaS guides | Six months later, if the MVP succeeded |
GST at 18 percent applies to the development invoice and is claimable as input credit if you are registered.
How to read an MVP quote
Six checks, ten minutes.
- Ask which rung it prices. If the proposal has admin roles, analytics and notifications in it, it is version 1.0 with an MVP label.
- Ask what question it answers, and what result would make you stop. A quote that cannot answer this is a feature list.
- Ask what has been left out, and why. A good agency volunteers the cut list. A bad one quotes everything you said.
- Ask for the instrumentation line. If nothing in the estimate measures the outcome, the MVP cannot report back.
- Ask how changes are handled when week four teaches you something. Fixed price with no change mechanism is a fight waiting to happen.
- Ask who owns the code and the accounts from day one. You, in writing.
The longer version, for choosing the company rather than reading the number, is in our guide to hiring a development company.
How to spend less and learn more
- Cut to one user type. The other side can be you, manually, for the first fifty transactions.
- Do the landing page first, even if you intend to build. A week and 50,000 rupees can save six months.
- Keep it web. Add the app when people are already using the web version enough to want one.
- Use the no-code tool for the plumbing and build only the part that is genuinely yours.
- Write down what would make you stop before the build starts. It is free, and it is the only thing that makes the result readable.
- Budget the month after launch. The MVP's value arrives in that month, not on launch day.
Questions people ask
How much does it cost to build an MVP in India?
Fifty thousand to 1.5 lakh for a landing page test, 1 to 3 lakh for a clickable prototype, 3 to 8 lakh for a working product with one real feature, and 5 to 25 lakh for a sellable first version. The worked estimate above, for a booking MVP with accounts and one core flow, comes to 47 person-days and about 7.5 lakh.
Why do MVP quotes vary from 2 lakh to 25 lakh?
Because the word covers four different things. A quote at the bottom is usually a prototype or a no-code assembly; one at the top is a first version with billing, admin and a second user type, which is a real product rather than a minimum one. Ask which rung is being priced before comparing two numbers.
How long does an MVP take?
Eight to twelve weeks for the third rung with an experienced team, one to two weeks for a landing page test, and three to four weeks for a clickable prototype. The week-by-week breakdown, and what actually stretches it, is in how long it takes to build an MVP.
Can I build an MVP with no-code?
Often, and it is usually the right first move. No-code stops working when your logic is unusual, when the data is sensitive, or when the tool's per-user or per-record pricing overtakes what owning it would cost. For testing demand and flow it is hard to beat.
What should an MVP not include?
An admin panel with roles, a second language, a notification system, analytics dashboards, a mobile app where web would do, and anything built for scale you do not have. Each is a real need for a real product, and each is cheaper to add once the product has earned it.
What does an MVP cost to run each month?
Three thousand to 25,000 rupees for hosting, transactional email, SMS and payment gateway fees at early volumes. Budget separately for the month after launch, which typically costs 15 to 25 percent of the build and is when the MVP actually does its job.
Should the MVP be an app or a website?
A website, in almost every case, because an install is a barrier for software nobody has used yet. Build the app when web usage is high enough that people ask for one, or when the product genuinely needs the camera, location or offline access.
Is a cheap MVP a false economy?
Only if cheap meant skipping the unhappy paths, the instrumentation or your own involvement. Cheap achieved by narrowing the scope to one question is the entire point. Cheap achieved by building the same scope badly is the false economy, and the two are easy to tell apart in a quote: one has a shorter feature list, the other has the same list with fewer days against it.
Getting a real number
Ranges end where your actual question begins. We scope MVPs backwards from the decision you are trying to make, which sometimes means we quote for less than you asked for, and occasionally means we tell you not to build yet. What that process looks like is on our MVP development page, and if the MVP is AI-shaped, on our AI MVP development page.