Xendit vs Midtrans: Which Payment Gateway to Pick in Indonesia

Every Indonesian app I ship eventually hits the same fork in the road: Xendit or Midtrans. They are the two payment gateways that actually matter here, and on paper they look almost identical — both do Virtual Account, QRIS, e-wallets, cards, and retail outlets. The real differences only show up once you are deep in the code, reading webhooks at 2am and reconciling settlements. This post is the comparison I wish I had before wiring up either one on a recent project for an Indonesian client.

I will keep this grounded in what a backend engineer cares about: which payment methods each one covers, how the fees actually stack up, what the developer experience and docs feel like, whether you can pay money out (disbursement), and finally a concrete rule for choosing between them for a startup versus an enterprise. Fees change often, so treat the exact numbers as a snapshot from the official pricing pages linked at the end and always re-check before you sign.

Payment method coverage

For the core Indonesian payment methods, coverage is close to a tie. Both gateways give you bank Virtual Accounts across BCA, BNI, BRI, Mandiri, Permata and CIMB, universal QRIS that any e-money app can scan, direct e-wallet channels, domestic and international cards, and retail over-the-counter payments at Alfamart and Indomaret. The one channel that genuinely differs is GoPay.

• Virtual Account — both support closed and open VA across the major banks; this is still the most-used method for Indonesian checkout.
• QRIS — identical on both, since QRIS is a Bank Indonesia standard; one QR accepts OVO, DANA, GoPay, ShopeePay, LinkAja and every bank app.
• E-wallets — both expose OVO, DANA, and ShopeePay as direct redirect or deeplink channels beyond the QRIS scan.
• GoPay — Midtrans is part of the GoTo group, so it offers a native GoPay app hand-off; on Xendit, GoPay users pay through universal QRIS instead.
• Cards and PayLater — both do Visa/Mastercard/JCB plus BNPL like Kredivo and Akulaku.

Fee structure compared

Both gateways charge zero setup and zero monthly fees — you only pay per successful transaction. The headline rates are close because a lot of Indonesian pricing is regulated or set by the underlying channel. QRIS, for example, is capped by Bank Indonesia at 0.7% for regular merchants, so it is the same everywhere. The table below is the snapshot I work from; the important nuance is VAT (PPN, currently 11%), which is added on top of most fees except QRIS, GoPay, and ShopeePay where it is already baked in.

Developer experience and docs

This is where Xendit earns its reputation. Its API is clean, REST-first, and consistent — you create an invoice or a specific payment-channel object, you get back a well-shaped JSON, and the webhook payloads are predictable. The docs read like they were written by engineers, with copy-paste examples and official SDKs for Node, PHP, Python, Java, and Go. If you are building a custom checkout from scratch, Xendit is simply the more pleasant place to live.

Midtrans is not bad — its Snap drop-in is genuinely fast to ship, and it has the widest set of ready-made CMS and framework plugins (WooCommerce, Laravel packages, and more). The trade-off is that the API surface feels older and more fragmented across its Core API, Snap, and Iris products, and the webhook signature verification trips up first-timers. Here is a minimal Snap token request to show the shape:

// Midtrans Snap: create a transaction token (Node.js)
const midtransClient = require('midtrans-client');

const snap = new midtransClient.Snap({
  isProduction: false,
  serverKey: process.env.MIDTRANS_SERVER_KEY,
});

const token = await snap.createTransaction({
  transaction_details: {
    order_id: 'ORDER-101',
    gross_amount: 150000, // IDR
  },
  credit_card: { secure: true },
});
// token.redirect_url -> open Snap checkout

# Xendit: create an invoice (any channel) via REST
curl https://api.xendit.co/v2/invoices \
  -u xnd_production_XXXX: \
  -H 'Content-Type: application/json' \
  -d '{
    "external_id": "order-101",
    "amount": 150000,
    "payer_email": "[email protected]",
    "description": "Order 101"
  }'
# Returns invoice_url + all enabled payment channels

Payouts and disbursement

This is the single biggest architectural difference. If your product needs to send money out — paying sellers on a marketplace, settling drivers, paying freelancers, or processing bulk refunds — Xendit treats disbursement as a first-class product with an API that pays out to 450+ banks and e-wallets, including single and batch payouts. Midtrans historically offered payouts through a separate product (Iris), but disbursement is not its center of gravity the way it is for Xendit. On a recent marketplace build I chose Xendit specifically because collecting and paying out through one provider kept reconciliation sane.

# Xendit: disburse money to a bank account
curl https://api.xendit.co/disbursements \
  -u xnd_production_XXXX: \
  -H 'Content-Type: application/json' \
  -d '{
    "external_id": "payout-seller-88",
    "bank_code": "BCA",
    "account_holder_name": "BUDI SANTOSO",
    "account_number": "1234567890",
    "amount": 500000
  }'

Which to pick: startup vs enterprise

There is no universally better gateway — there is only the right fit for what you are building and how fast you need to move. Here is the decision rule I actually use.

• Early-stage startup shipping fast on a CMS or framework with existing plugins — Midtrans, for time-to-live and the native GoPay hand-off Indonesian buyers expect.
• Custom product on a modern backend where API quality matters — Xendit, for the cleaner API and docs your team will maintain for years.
• Marketplace, fintech, or any model that pays money out — Xendit, because collect plus disburse in one provider is a huge operational win.
• Regional ambition beyond Indonesia — Xendit also operates in the Philippines and wider Southeast Asia, so one integration scales across borders.
• Enterprise needing named support, custom rates, and hardened uptime — negotiate both; the winner is whoever gives your finance team the best rate card and SLA.