How long should an MVP take to build?

Why estimates vary so much

"MVP" means very different things to different people. For one founder it's a single booking journey with sign-in and payments. For another it's a multi-sided marketplace with an admin portal. Both get called an MVP, and they're not the same size of job.

The honest way to estimate is to agree what the product must prove first, then size only that. Everything else goes on a list for later.

What actually decides the timeline

User journeys. Each distinct thing a user can do adds build, testing and design time.
Integrations. Payments, sign-in providers and third-party APIs each come with their own edge cases.
Back office. Admin screens, reporting and exports are often forgotten, and often essential.
Data sensitivity. Personal or financial data raises the bar for security and compliance.
Decision speed. A week waiting for sign-off is a week added, however fast the engineering is.

Why agentic delivery is faster, without skipping steps

An agentic MVP still goes through discovery, sprint planning, build, testing and review. What changes is the time each step takes. Agents handle research, implementation, tests and refactoring under supervision, while a senior engineer makes the architecture decisions and reviews every change before it merges.

We don't remove the steps. We remove the waiting between them. You can read more in how agentic development works.

How to keep an MVP to weeks, not months

Write down the one thing the MVP must prove.
Cut every journey that doesn't help prove it.
Build one journey end to end in the first sprint.
Put real users on it as early as possible.
Agree who makes decisions, and how fast.