Kenya vs South Africa payments: what builders need to know about both

Where the money sits, and who runs the rails it moves on. The Kenyan guides on this site start at an operator's API: Daraja is Safaricom's developer portal for its APIs^{[Safaricom Daraja developer portal, checked 2026-11-05]}, and it lets developers create a sandbox app and test it^{[Safaricom Daraja developer portal, checked 2026-11-05]}, and our M-Pesa hub covers it from first sandbox call to production. One company, one API, one wallet to reach.

South Africa is built around bank accounts and schemes the banks share. PayInc runs the retail payment schemes: EFT credits and debits, DebiCheck, Registered Mandate, PayShap and real-time clearing^{[SARB media release and FAQ, checked 2026-11-05]}. A business that is not a bank reaches those schemes through one: as the SARB Governor put it in October 2026, at the moment anyone who wants to access the payment system has to go through a bank^{[BIS: Kganyago, checked 2026-11-05]}. And the Kenyan model was tried there: Vodacom announced on 9 May 2016 that M-Pesa in South Africa would close on 30 June 2016^{[Vodacom media release, checked 2026-11-05]}.

The verdict for builders is simple to state and easy to forget. In Kenya, design around the wallet and the operator's API. In South Africa, design around bank-based methods reached through a payment provider. The same team can do both well, but not with the same assumptions.

This section of the hub is the bridge between our Kenyan guides and the Southern Africa hub. For the regional picture beyond South Africa, see payments in Southern Africa.

Because South Africans already had bank accounts, and mobile money needs a crowd to work. The history is short and well documented:

• Vodacom and Nedbank launched M-Pesa in South Africa on 31 August 2010^{[Vodacom media release, checked 2026-11-04]}.
• Vodacom's 2015 integrated report counted one million registered M-Pesa customers in South Africa, of whom 76,000 were actively using the service^{[Vodacom, checked 2026-11-05]}.
• Vodacom announced on 9 May 2016 that the service would close on 30 June 2016^{[Vodacom media release, checked 2026-11-05]}.

Vodacom explained the decision plainly: Shameel Joosub said the business sustainability of M-Pesa is predicated on achieving a critical mass of users, and that high levels of financial inclusion in South Africa left little prospect of achieving it in its current format in the mid-term^{[Vodacom media release, checked 2026-11-05]}. A later study added other factors: FinMark Trust's 2017 report pointed to agent network challenges, poor technology choices, an institutional environment that constrained the services, and established competition in domestic remittances^{[FinMark Trust, checked 2026-11-05]}.

Vodacom was not alone. MTN also decided in 2016 to decommission its mobile money offering^{[MTN SA press release, checked 2026-11-05]}, and MTN came back later: on 18 December 2019 MTN announced that MoMo would go live in South Africa in January 2020^{[MTN Group media release, checked 2026-11-05]}, initially for MTN customers only^{[MTN media release, checked 2026-11-05]}.

The lesson for a product team is that rails follow where people keep their money. The full story is in why M-Pesa failed in South Africa, and the comparison with markets where M-Pesa took a different path is in M-Pesa Kenya vs Mozambique vs South Africa.

Cards, and a growing set of bank-based methods. PayShap has scale, with more than 1 billion transactions worth more than R1 trillion since launch^{[BIS: Kganyago, checked 2026-11-05]}, but it is small next to cards: last year PayShap processed only about 6% as many transactions as cards, and under 40% of EFT credit volumes^{[BIS: Kganyago, checked 2026-11-04]}.

A consumer survey shows how wide the menu has become. Stitch's 2026 Consumer Payments Report found that, among new payment methods consumers tried in the past 12 months, one-click wallets reached 58%, BNPL 39%, Capitec Pay 37% and PayShap 35%^{[Stitch blog, checked 2026-11-05]}. Capitec Pay matters because of who can use it: Capitec has more than 26 million active clients, its FY2026 results say^{[Capitec results, checked 2026-11-04]}.

For a Kenyan team used to building around M-Pesa, the South African checkout is a portfolio:

• Cards carry far more transactions than PayShap, so they are not optional.
• Pay-by-bank methods such as Capitec Pay and instant EFT serve customers who would rather not use a card; PayShap vs instant EFT vs Capitec Pay compares them.
• PayShap Request will feel the most familiar to a Daraja developer; what PayShap is and how it works covers the scheme.
• Buy now, pay later reaches many shoppers, as the survey shows, and is a separate integration with its own rules.

Payment gateways in South Africa compares the providers that bundle these methods. If you would rather have the South African side of a Kenyan product scoped and built for you, Mctaba Labs payment integration does that work.

On the Kenyan side you integrate the rail. You register on Safaricom's portal, use the sandbox, then apply for production; the Daraja API tutorial for beginners and Daraja vs aggregators cover the choice between going direct and using a gateway.

On the South African side you integrate a provider, and there are many to choose from: the SARB's 2025/26 Annual Report counts 39 banks and 486 non-bank payment service providers in South Africa's payments ecosystem in calendar 2025^{[SARB Annual Report 2025/26, checked 2026-11-05]}. One name will be familiar: Paystack has operated in South Africa since 2021^{[Paystack blog, checked 2026-11-05]}. A team that knows Paystack from Kenya should still check which methods it offers in each country before planning around it. Payfast vs Paystack vs Ozow vs Peach Payments compares the main options.

One gotcha catches e-commerce teams arriving from elsewhere: Shopify Payments is not offered in South Africa^{[Shopify Help Center, checked 2026-11-04]}, so a Shopify store there needs a third-party gateway from the start.

The API-level detail, STK Push against PayShap Request, payer identifiers, callbacks and limits, is in Daraja vs PayShap.

Both countries have an instant bank rail beside the methods most people talk about. Pesalink has provided real-time bank transfers in Kenya since November 2016^{[Pesalink, checked 2026-11-05]}, and in South Africa PayShap launched on 13 March 2023^{[SARB media statement, checked 2026-11-05]} as the instant rail between bank accounts. Both let a business move money between banks in real time; they differ in who runs them, how a payer is addressed and how a developer gets access.

On PayShap, you can pay to an alias instead of a bank account number^{[SARB retail CBDC background note, checked 2026-11-04]}, which brings it closer to the phone-number experience Kenyans know from M-Pesa. Pesalink vs PayShap compares the two in detail, and our existing guide to Pesalink bank transfers through Paystack covers the Kenyan access route.

In South Africa, the central bank, directly and recently. PASA's recognition as the payment system management body was withdrawn, and its functions transferred to the SARB and PayInc, from 2 September 2026^{[SARB, checked 2026-11-05]}, and authorisation and registration of system operators and third-party payment providers moved to the SARB on 11 August 2026^{[SARB, checked 2026-11-04]}. The next change is already drafted: the SARB published the draft directive for payment activities, the Authorisation Framework, in November 2025^{[SARB Annual Report 2025/26, checked 2026-11-05]}.

What that means for a Kenyan company entering South Africa:

• Collecting money on behalf of others brings in third-party payment provider rules; read what a TPPP is before you design the money flow.
• New payment products should be scoped against the non-bank Authorisation Framework, not the older registration model.
• The governance change itself is explained in what happened to PASA.

On the Kenyan side, the rules a developer meets are covered in our guide to CBK payment service provider rules developers should know.

South Africa has formal machinery for pulling money from bank accounts, and its dispute rules have just changed. Debit orders and DebiCheck are the established collection rails, and the dispute period for debit orders, DebiCheck and Registered Mandates fell from 365 days to 60 calendar days from 13 April 2026^{[PASA, checked 2026-11-05]}. A newer option sits on top of a bank app: Stitch announced Capitec Pay VRP on 4 December 2025^{[Stitch blog, checked 2026-11-05]}, letting customers pre-approve variable charges.

So a Kenyan SaaS team cannot simply port its billing logic. Recurring payments in South Africa compares card tokenisation, DebiCheck and Capitec Pay VRP, and for the Kenyan side our guide to building a Kenyan SaaS subscription billing system shows the patterns there.

Through cross-border rails and the controls that sit on them, not through a shared domestic scheme. On the South African side, authorised dealers may only make foreign currency payments for imports against the documents the Currency and Exchanges Manual lists^{[SARB, checked 2026-11-05]}, and the same discipline applies when money leaves for Kenya.

If you are a South African business with Kenyan customers, accepting payments from Kenyan customers covers the collection side. For corridors, costs and who runs what across the region, see cross-border payments in Southern Africa.

Pick the page that matches the move you are making:

You are	Read first
A Kenyan team adding South Africa	What Kenyan fintech builders get wrong about South Africa
A South African team adding mobile money markets	What South African teams get wrong about mobile money
A developer comparing the two main APIs	Daraja vs PayShap
Building one product for both markets	One codebase, two rails
Comparing the instant bank rails	Pesalink vs PayShap
Weighing feature-phone channels	USSD in Kenya and South Africa

Five decisions, made early, save a rebuild later:

1. Choose rails per market, not per product. Daraja on the Kenyan side; a provider with cards and at least one pay-by-bank method on the South African side.
2. Model the payment request, not the rail. One request record and one status model, with an adapter for each rail, keeps order code clean when you add a third market.
3. Type your payer identifiers. A phone number, a bank account and an alias are different things with different validation. Store them as such.
4. Keep currencies and ledgers apart. Shillings and rand settle at different times through different parties; reconcile each market on its own reports before you consolidate.
5. Map compliance per market. South Africa's FICA, KYC and POPIA rules for payment products and the Kenyan rules each need their own owner on your team.

The engineering habits Kenyan developers learned on Daraja carry straight over: asynchronous flows, callbacks confirmed before fulfilment, idempotent handlers and daily reconciliation. What does not carry over is the idea that a single payment method can cover a market.

For the full regional view, country by country, see payments in Southern Africa. And when you want one payment layer built across both markets, Mctaba Labs payment integration builds integrations on both sides.