Payment processing on an MVP costs two things: a per-transaction fee of roughly 2-6% depending on which processor you pick, and 3 days to 4 weeks of engineering time depending on whether you use a hosted checkout or build one yourself. The fee line isn't where the real cost hides.
Most founders ask "which payment processor is cheapest" and get the wrong answer, because the fee line isn't where the cost lives. The cost lives in what happens the first time a customer in a country you didn't plan for tries to pay, or the first time you owe VAT in a jurisdiction you've never registered in. Stripe and Razorpay charge the lowest transaction fees but leave you responsible for tax compliance; Paddle and Lemon Squeezy charge more per transaction but act as merchant of record, which removes the tax problem entirely — the right pick depends on whether you're selling to India, the US, or both.
Which payment processor is cheapest for an MVP?
By transaction fee alone: Razorpay, at 2% for domestic Indian cards and UPI, with no setup or monthly fee. Stripe is close behind at 2.9% + $0.30 for US card-present online payments, rising to 3.9% plus a 1% currency-conversion fee for international cards. Both are gateways, not merchants of record — you collect the money, and you own any tax registration the sale triggers.
Paddle and Lemon Squeezy charge more, typically 5% + $0.50 per transaction. That premium buys merchant-of-record status: they are the legal seller, they collect and remit VAT/GST/sales tax in every jurisdiction automatically, and you invoice them once instead of registering for tax in 40 countries. For a US-only B2B SaaS with a handful of enterprise customers, that premium is usually not worth it. For a self-serve product selling globally from day one, it often is.
How long does Stripe integration actually take?
Stripe Checkout — the hosted, pre-built payment page — is a 3-to-5-day integration for a standard subscription MVP: create the product and price objects, redirect to Checkout, handle the checkout.session.completed webhook, and provision access. That's the version most MVPs should ship first.
Custom Stripe Elements — building your own payment form inside your app instead of redirecting to Stripe's hosted page — is a 2-to-4-week job once you add the real requirements: webhook handling for failed payments and disputes, subscription upgrade/downgrade proration, dunning emails for failed renewals, and a customer billing portal. We've built both inside the same MVP timeline more than once, and the decision is almost always the same: ship Checkout in week one, only justify custom Elements once a specific enterprise buyer asks for a branded checkout page you can't get any other way.
What does merchant-of-record actually save you?
It saves you from having to know that Germany's VAT rate is 19%, that some US states require sales tax on SaaS and some don't, and that if you cross certain revenue thresholds in a country you may owe a local tax filing whether or not you have an entity there. A merchant of record takes that risk and that paperwork off your plate for a fixed markup on every transaction.
The trade-off is control: you don't own the customer relationship on the payment side, refund and dispute policies run through the MoR's rules, and payout timing is on their schedule, not yours. For an MVP with under a few hundred transactions a month, that trade is usually worth it — the 3-point fee premium is smaller than the cost of one afternoon spent figuring out VAT registration.
Comparison: which processor for which MVP
| Processor | Fee | Who's the merchant | Best for |
|---|---|---|---|
| Razorpay | ~2% domestic, ~3% international | You | India-first products billing INR |
| Stripe | 2.9% + $0.30 (US), 3.9% + 1% intl | You | US-first SaaS, subscriptions, marketplaces |
| Paddle | ~5% + $0.50 | Paddle | Global self-serve SaaS avoiding tax registration |
| Lemon Squeezy | ~5% + $0.50 | Lemon Squeezy (now part of Stripe) | Solo/small-team products selling digital goods globally |
What does an India-built, US-selling MVP need for payments?
This is the geo case we see most often: an Indian founder or an Indian-based team building for US customers. The clean setup is Stripe for the US-facing billing (it doesn't require a US entity to start, though payouts and tax reporting get simpler once one exists), and Razorpay only if there's a genuine India-side customer base paying in INR. Running both from day one is over-engineering for an MVP with zero customers — pick the market you're launching into first, wire that processor, and add the second only once revenue in that market is real. We cover the broader stack decision, including where billing sits relative to auth and hosting, in our opinionated stack for shipping an MVP in 6-16 weeks.
What does this add to the MVP budget?
Hosted checkout (Stripe or Razorpay, no custom UI) adds roughly 3-5 days to a build — a rounding error against a 6-16 week MVP timeline, covered in what an MVP costs in 2026. Custom Elements with full subscription lifecycle management adds 2-4 weeks, which is a real scope decision, not a checkbox — it belongs in the same conversation as every other cut we walk through in how to scope an MVP. Post-launch, expect a few hours a month of maintenance as card networks update rules and webhook payloads shift — the kind of small, recurring work we detail in what a post-MVP retainer costs.
The default we ship
Stripe Checkout for US-billing MVPs, Razorpay Checkout for India-billing MVPs, and we don't reach for Paddle or Lemon Squeezy until a client's product is genuinely global self-serve with no sales team touching the deal. We build the webhook handler and the billing portal before launch, not after the first failed-payment support ticket. Payments are the one part of an MVP where the "we'll fix it later" instinct is expensive to reverse — a checkout page is easy to swap, a tax filing you skipped is not.
The question worth asking your dev shop before they write a line of billing code: "Are we the merchant of record, or is the processor?" If they don't have an immediate answer, that's the answer.
Written 2026-09-25 by Naman Barkiya.