Breaking Ground: Eyooly’s Expansion Blueprint for Equatorial Guinea and Beyond

An Oil Nation Waking Up to Digital

There is a particular kind of opportunity that only reveals itself when you look past the headline numbers. Equatorial Guinea produces roughly 100,000 barrels of oil per day. It has a GDP per capita that, on paper, places it among the wealthiest nations in sub-Saharan Africa. And yet, walk through the markets of Malabo or Bata and you will find something that should not exist alongside those numbers: a commerce infrastructure that is almost entirely offline, fragmented, and opaque.

That gap is not a failure. It is an invitation.

Eyooly launched in Equatorial Guinea with a straightforward thesis: local commerce, mobility, and delivery should work as reliably as calling a friend. Not because the market demanded a sleek interface, but because the absence of a trusted, organized digital layer was the single biggest friction point for ordinary buyers, sellers, and riders trying to do ordinary things every day.

We are now past the question of whether this market wants what we are building. The platform has answered that. What this piece is about is how we scale it  technically, operationally, and across borders without breaking the trust we have spent this early phase building.

Reading the Market Honestly

Before strategy, there must be honesty about what the numbers actually say.

The Fintech Times has documented what anyone working in this region already knows: Equatorial Guinea’s digital financial ecosystem is underdeveloped, with fewer than five to ten active providers across the entire fintech stack. That is not a competitive landscape that is a near-empty field. But it is an empty field with a fence around it, and understanding that fence matters more than the open space.

Regulation here flows through the BEAC the Banque des États de l’Afrique Centrale which governs financial service licensing across the CEMAC zone. That is not an obstacle to be routed around. It is a constraint to be built with. Any platform operating payments or wallet infrastructure in this region that is not thinking about BEAC compliance from day one is building on sand. Our architecture reflects this: payment logic is isolated in its own module, commission and settlement flows are auditable by design, and the wallet layer (Eyooly Pay) is sequenced after our core marketplace proves traction precisely so that when we approach licensing, we arrive with evidence, not projections.

On connectivity, Data Reportal puts internet penetration in Equatorial Guinea at 60.4%, with approximately 1.18 million users online. That figure matters less as a ceiling and more as a calibration. It tells you that the mobile-first assumption is correct this is not a desktop market, it never will be — and that the gap between “connected” and “transacting digitally” is still very wide. That gap is where Eyooly operates.

The World Bank’s Digital Agenda for Equatorial Guinea (ADIGE) provides the most important macro signal: the government has an explicit policy mandate to diversify the economy away from oil dependency through digital infrastructure investment. That is not marketing language borrowed from a development report. That is fiscal reality. When a government that earns the majority of its revenue from a depleting resource starts writing digital transformation policy, the question is not if digital commerce scales it is who builds the rails.

The Proof-of-Work: What We Have Already Built

Let me be direct about what the current platform metrics represent, because I think the framing matters.

Eyooly today has over 2,400 active ads on the marketplace, has facilitated more than 1,200 deliveries, operates across 5 cities, and has onboarded over 800 sellers. Users log in without passwords a deliberate security and UX decision I will explain in a moment.

Some observers will look at those numbers and call them early-stage. They are right. But they are describing the wrong thing.

What those numbers actually represent is a verified feedback loop. We know how a seller in Malabo behaves when listing a product for the first time. We know where delivery confirmation breaks down. We know which categories attract the most contact requests and which friction points cause users to abandon a session. We know this because we have real transactions, real disputes, and real support interactions not user interviews, not focus groups, not assumptions imported from a Lagos or Nairobi playbook.

That is proof-of-work. Every architectural decision we make from here is grounded in observed behavior, not modeled behavior. That distinction is the most valuable technical asset we have.

The Technical Architecture for Scale

I want to explain how we are building this, because the choices are not obvious and they are not accidental.

Modular Monolith First

The platform runs on a FastAPI backend structured as a modular monolith one deployable unit with clear internal domain boundaries. We have explicit modules for the marketplace, logistics, food, dispatch, trust, and admin functions. Each module enforces its own data ownership and API contracts internally.

This is a deliberate departure from the microservices-first architecture that became fashionable in the mid-2010s among platforms with very different operational profiles than ours. Microservices distribute operational complexity across infrastructure. For a team in our stage moving fast, in a market where infrastructure reliability cannot be taken for granted that distribution is a liability, not an asset.

When we are processing 10,000 daily deliveries across three countries, we will revisit this. Until then, the monolith is the honest choice. Anyone who tells you otherwise is optimizing for architectural elegance over operational reality.

Location Intelligence as Core Infrastructure

PostGIS — the geospatial extension for PostgreSQL is not a feature bolt-on in our stack. It is load-bearing infrastructure. In a market where formal addresses are inconsistent or absent, where delivery zones are informal, and where driver availability is distributed across a city without organized hubs, geospatial querying is the difference between a dispatch system that works and one that guesses.

Every delivery task, service zone, driver location update, and price rule in Eyooly Go runs through PostGIS primitives. That means we can query: which available drivers are within 2km of this pickup point, ranked by distance, filtered by service type and operating zone. In real-time. Without approximation.

This is not a technical detail — it is the operational foundation of our logistics reliability.

Passwordless Authentication and Why It Matters Here

The decision to offer passwordless login on Eyooly is not primarily a UX choice. It is a security and inclusion choice with specific relevance to this market.

Password-based systems create predictable failure modes in low-literacy and multi-device environments forgotten credentials, shared phones, reused passwords that become breach vectors. The “sin contraseñas” approach login via OTP, magic link, or social auth eliminates the most common account compromise vector while simultaneously lowering the barrier to entry for first-time digital commerce users.

As an AI and Cybersecurity Strategist, I will say plainly: the most dangerous assumption in African fintech expansion is that the threat model is identical to Europe or North America. It is not. SIM-swap fraud, account takeover via social engineering, and marketplace fraud targeting new users are the dominant risk vectors here. Our authentication architecture, combined with internal user UUIDs that are decoupled from any external identity provider, gives us the ability to rotate credentials, flag risk signals, and audit account activity without being dependent on a single point of identity failure.

Secure payments across borders start with secure identity at the account level. That is first principles, not feature development.

The AI Layer: Present but Not Critical Path

Eyooly uses AI for listing category suggestions, semantic product search, moderation support, and seller onboarding assistance. Groq and Jina AI are in the stack for inference speed and embedding quality respectively.

But nothing in our critical path login, order creation, delivery assignment, payment settlement touches an AI endpoint. This is intentional. AI credit exhaustion, model API downtime, or inference latency spikes cannot be allowed to block a delivery confirmation or a payment. The platform must work completely without AI. AI makes it faster and smarter. It does not make it run.

Adams Cosmas is a Tech Founder, Systems Engineer, and AI & Cybersecurity Strategist. He is the author behind Olybee and a contributor to Eyooly’s strategic and technical direction.

Eyooly is a product of OLYTECH Design Ltd (RC: 1840343). Learn more at eyooly.com · @eyoooly