Offline Learning Solutions in Sub-Saharan Africa

Why offline still matters

Even where mobile data exists, costs of 1–5 USD per GB price most rural classrooms out of streaming video, textbook downloads, or AI tutors. Power is intermittent. Bandwidth is shared. Offline-first delivery is not a workaround — it is the primary channel.

Two distribution models

Every offline learning project falls into one of two patterns.

1. Physical distribution (SD cards, USB drives)

A curriculum is preloaded onto storage media and physically delivered to each student or device. EduCard uses this model: a complete K-12 curriculum and an on-device AI tutor ship on an $8 SD card. There is no server, no network, and no recurring cost per classroom.

• Strengths: zero infrastructure, per-student cost is the cost of the card, content runs on any phone or tablet with an SD slot.
• Weaknesses: content updates require re-distribution, no central analytics, harder to enforce one shared "class" view.

2. Digital distribution (local servers — Kolibri, RACHEL, Raspberry Pi)

A small local server (often a Raspberry Pi) hosts content like Khan Academy Lite, Wikipedia, or open textbooks, and student devices connect over local Wi-Fi. Kolibri (Learning Equality) is the most widely deployed example and powers a large share of school-server projects globally.

• Strengths: one update reaches the whole school, supports class-level analytics, students share a single "library."
• Weaknesses: requires a server, power for that server, Wi-Fi range, and someone trained to maintain it.

How to choose

The right model depends on three honest questions about the school.

• Power. If the school has no reliable electricity for a server, physical SD cards win — they sip power on the student's device only.
• Devices. If most students already share a phone or tablet at home, SD cards reach them outside the classroom. If devices are pooled at school, a server makes sense.
• Maintenance. If no on-site technician exists, avoid solutions that require server administration. SD cards have nothing to break.

EduCard's approach

EduCard targets the lowest-resource case: no internet, no server, sometimes no grid power. A single $8 SD card delivers the full K-12 curriculum plus an AI tutor that runs locally on the student's phone. We chose physical distribution because the marginal cost of reaching one more child is the cost of one more card — nothing else.

We see Kolibri and similar server-based projects as complementary, not competing. A district with reliable power at the school and intermittent connectivity at home benefits from both: a Kolibri server in the classroom and EduCard SD cards for at-home study.

Further reading

• Learning Equality — Kolibri (learningequality.org)
• World Possible — RACHEL (worldpossible.org)
• African Union — Digital Education Strategy 2023–2028