Every founder asks this question before they ask anything else, and every agency gives a vague answer before they've even heard the project. Here's a straight one, based on what projects of each type actually run in India in 2026 — not a sales number designed to get you on a call.
The short version
- A focused MVP (one core workflow, one platform): roughly ₹6–18 lakh ($7K–$22K), 6–10 weeks.
- A full product (web + mobile, auth, payments, admin panel): roughly ₹25–70 lakh ($30K–$85K), 3–6 months.
- Enterprise integration or internal tooling (connecting to existing systems, migrating legacy data): roughly ₹40 lakh–₹1.5 crore ($48K–$180K+), highly variable.
- AI/automation add-on to an existing product: roughly ₹8–30 lakh ($10K–$36K), depending on how much custom model work is involved versus wiring up existing APIs.
These are ranges for senior-led, India-based teams — not the lowest bids you'll find on freelance marketplaces, and not Silicon Valley agency rates either. If a quote is wildly below this range, ask who's actually writing the code; if it's wildly above, ask what you're paying for beyond the engineering.
What actually moves the number
1. How well-defined the scope is before you start
The single biggest cost driver isn't the tech stack — it's ambiguity. "Build me an app like Uber but for X" has no fixed price because it has no fixed scope. A one-page spec with concrete user flows, a clear list of what's out of scope, and acceptance criteria for "done" routinely cuts 20-30% off both cost and timeline, because the team isn't re-negotiating scope mid-build.
2. Integrations, not features
A login screen is cheap. A login screen that has to sync with your existing HR system, respect three different data-residency rules, and handle SSO for five different client organizations is not. When you're scoping, the integrations list is usually the real cost driver — features are mostly templated at this point; integrations almost never are.
3. Who's actually building it
A team of 2 senior engineers who've shipped production systems at scale will usually out-deliver a team of 6 junior developers, for less total cost and in less time — fewer handoffs, fewer bugs that need a second pass, no one learning the stack on your dollar. This is the single most common thing founders get wrong when comparing quotes: a lower day-rate with more people on the team is very often a higher total cost.
4. Fixed-scope vs. time-and-materials
Fixed-scope pricing sounds safer, but it only works when the scope is genuinely fixed — which means it only works well for the well-defined projects in point #1. For anything exploratory (a new AI feature, a product pivot), time-and-materials with weekly visibility into what's being built is usually cheaper in practice, because you're not paying a risk premium the vendor baked in to cover scope creep they expect.
Two mistakes that blow up budgets
Mistake 1: Treating the first version as the only version
Teams that try to build "the whole product" before shipping anything spend 2-3x more before they get real user feedback — and often discover the feature they spent the most on isn't the one users actually want. Scoping a genuine MVP (not a watered-down full product, an actual minimum viable slice) is the highest-leverage cost decision you'll make.
Mistake 2: Skipping the discovery phase to "save time"
A 1-2 week paid discovery/scoping phase feels like overhead when you're eager to start building. In practice, it's the cheapest insurance in the entire project — it's where integration surprises, unclear requirements, and unrealistic timelines get caught before they're expensive to fix instead of after.
A simple way to sanity-check any quote
Ask three questions: Who specifically will be writing the code (names, not just "our team")? What's explicitly out of scope? What happens if the scope changes in week 3? If an agency can't answer these clearly, the number they gave you isn't really a quote — it's a guess with a decimal point.