Every African startup that processes payments eventually builds the same infrastructure. A Chapa integration here, a Telebirr adapter there, a webhook retry system, a transaction reconciliation layer, a dashboard for the ops team. Six months and tens of thousands of dollars later, they have a fragile, under-documented payment stack that only one engineer truly understands.
We've built this stack four times across different companies. The fifth time, we decided to open-source it.
The problem
Africa has deeper payment fragmentation than anywhere else on Earth. Each country has its own set of mobile money providers, bank transfer networks, and payment gateways — each with its own API style, authentication flow, and settlement timeline.
In Ethiopia alone, you might need to integrate Telebirr (USSD push via Ethio Telecom), Chapa (hosted checkout with card support), CBEBirr (Commercial Bank of Ethiopia's mobile money), and EthSwitch (the national interbank switch). Each has a different sandbox, different credential format, and different error handling approach.
Move to Kenya, and you're dealing with M-Pesa's STK push, Airtel Money, and a completely different regulatory environment. Nigeria adds Paystack, Flutterwave, and their own set of complexity.
What Zirzir is
Zirzir is a three-layer open-source payment infrastructure stack:
The SDK abstracts away provider differences. A Chapa card payment and a Telebirr USSD push look identical to your code — one Transaction type, one set of statuses, one webhook format. Available in TypeScript and Python, with Go coming soon.
The Server adds operational infrastructure: a REST API, webhook engine with configurable retry schedules, transaction ledger with full audit trail, multi-project isolation, and a real-time dashboard. It's a single Go binary that deploys anywhere.
Zirzir Cloud is the managed version for teams that don't want to run infrastructure. Same capabilities, with SLA guarantees, automatic backups, and priority support.
Why open core
We chose the open-core model deliberately. The SDK and server are fully MIT-licensed. You can self-host the entire stack, fork it, modify it, and use it commercially without any restrictions.
Cloud is how we fund development. We'll never move features behind a paywall after you depend on them. If it's in the open-source version today, it stays there.
What's next
We're expanding provider coverage across the continent. M-Pesa and Airtel Money for East Africa are already in beta. Paystack and Flutterwave for Nigeria are next on the roadmap, followed by MTN MoMo for West and Central Africa.
The plugin architecture makes adding new providers straightforward — implement a Go interface, drop it into your deployment.
We believe African startups deserve better than rebuilding the same payment infrastructure from scratch. That's why we built Zirzir.