Almost every UAE product build reaches the same fork in week two, and it is rarely decided on merit. Someone picks the gateway their bank suggested, or the one a previous developer knew, and the consequences show up months later in settlement times and failed recurring charges.
The choice is about your entity, not the API
Every gateway on this list will take a card payment. What separates them is whether they will onboard your legal entity, which currencies they settle to your account in, and how long the money takes to arrive. Those are commercial questions that get treated as technical ones.
| Gateway | Fits | The catch |
|---|---|---|
| Network International | UAE entities selling to UAE customers in AED | Onboarding is bank-paced. Expect documents and weeks, not an afternoon. |
| Telr | SMEs and first products, faster to get live | Fewer options once you need unusual flows. |
| PayTabs | Selling across several Gulf markets from one integration | Regional breadth is the reason to choose it; if you only sell in the UAE it buys you little. |
| Stripe | Products selling internationally, or SaaS with subscriptions | Best developer experience by a distance, but check entity eligibility and settlement currency before you build on it. |
What the integration actually costs to build
The happy path — customer pays, you get a success callback — is a day of work on any of them. The rest is where the time goes, and it is the part that never appears in a demo.
- Webhook handling that is idempotent, so a retried notification does not credit an order twice
- Reconciliation between what the gateway says settled and what your database believes
- Refunds and partial refunds, including what that does to an invoice you already issued
- 3-D Secure flows and what the user sees when the bank declines
- Failed recurring payments: retry schedule, grace period, and when access is actually cut off
Backend and core work of this kind consumes a substantial share of a build budget and is invisible in a demo, which is exactly why cheap quotes cut it first.
The mistake worth avoiding
Do not let the gateway shape your data model. Store your own order and payment records, keyed by your own identifiers, and treat the gateway as a service you call. Products that wrote the gateway's assumptions into their database find that switching provider later is a rebuild rather than a swap, and in this region provider switches happen more often than anyone plans for.