Most Indian store owners end up choosing between these two, and the decision is usually made on a fee sheet. Fees matter, but the difference that shows up six months later is whether money only comes in, or also goes out to sellers, vendors or riders.
The short answer
- Razorpay if you are building a product, need subscriptions and want the largest feature surface.
- Cashfree if you also need to pay people out, or if settlement speed is the deciding factor.
Everything below is the reasoning behind those two lines.
Where they actually differ
| Razorpay | Cashfree | |
|---|---|---|
| Checkout | Hosted, embedded and direct API options | Hosted and API options with a similar split |
| UPI | Autopay mandates and intent flows | Autopay mandates and intent flows |
| Payouts | Available as a separate product | Payouts sit in the same account as collections |
| Subscriptions | Mature, with mandate management | Supported, with fewer surrounding tools |
| Multi-currency | Supported for international cards | Supported, with higher fees for cross-border |
| Best fit | SaaS, apps, product businesses | Marketplaces, payout-heavy businesses |
The payouts row is the one that changes project scope, because a payout integration needs beneficiaries, approvals and a transfer ledger whether or not the gateway makes it convenient.
Fees, without inventing numbers
Both providers price domestic cards and UPI differently, both charge GST on their fee, and both negotiate rates by volume. Published rates change, and a page quoting them goes stale.
Look at four things instead: the rate for your payment mix, the settlement cycle you are offered, the cross-border rate if you sell abroad, and the chargeback fee. Ask each provider for those four in writing, for your expected volume, before comparing.
UPI success rate is not a marketing number
Both gateways route UPI through the same underlying rails, and success rates move with the customer’s bank, the app they use and the time of day. A provider claiming a better rate is usually quoting a specific period.
What you control is the fallback: when a UPI payment fails or times out, does the checkout offer another method without losing the cart? That single screen matters more than a percentage point.
Integration effort
For a hosted checkout, the work is comparable: create the order before the gateway call, verify the return, handle the webhook, record refunds. The differences appear in three places:
- Signature verification. Each provider signs differently, and copying the verification code between them fails quietly.
- Payout APIs. Cashfree’s payout model is part of the same account, so a marketplace integrates one provider instead of two.
- Settlement reports. Formats differ, and the reconciliation import has to be written per provider.
We cost the standard collection integration from Rs 4,999 either way; the payout or marketplace module is where the two diverge. The pricing detail is on Cashfree integration cost and Razorpay integration cost.
When to choose Razorpay
- The product is a SaaS or app with recurring billing.
- You want payment links, invoices and subscriptions from one dashboard.
- The team wants the largest set of SDKs and examples, because the build will be self-served later.
- International cards are a small but real part of revenue.
When to choose Cashfree
- Money has to move both ways: seller settlements, vendor payouts, rider earnings.
- You are building a marketplace and do not want a second payout provider.
- Settlement speed is a competitive requirement, not a preference.
- The finance team wants collections and payouts reconciled from one source.
What we would not do
Do not run both as primary gateways in year one. Routing between two providers doubles the reconciliation work, and a small store rarely has the failure volume that justifies it.
Keep the second provider as a documented fallback for the day the first has an outage, and integrate it when the volume makes the complexity worth carrying.
The integration itself
Whichever you choose, the failure paths decide the quality of the build: a declined card, a timeout after debit, a duplicate webhook, a refund after settlement. Our payment gateway integration scope covers those, and the full gateway comparison covers PayU, Paytm, PhonePe and Instamojo if they are still in the running.
If the store itself is being built, start from ecommerce website development and bring the gateway decision into the first week, not the launch week.
