MeshLib: A Hybrid Offline Mesh Network for Academic Content Delivery in Connectivity-Constrained Nigerian Campuses

Abstract

Millions of tertiary students across Nigeria face a persistent barrier to academic resources: the cost of mobile data. Existing offline-first educational technology solutions are typically designed around either static content caching or single-device offline access, and rarely address the redundant, device-by-device downloading that occurs when many students in the same physical space independently retrieve identical study materials. This paper presents MeshLib, a two-stage system designed to close this gap through device-to-device content propagation. The first stage, an ESP32-based seed node prototype, was deployed and field-tested at a national hackathon, demonstrating that a single content download could propagate across several devices without additional data cost. The second stage, a Raspberry Pi-based architecture incorporating an offline artificial intelligence tutor, was designed to extend the system beyond static content delivery — a capability that exceeds the computational limits of the original ESP32 hardware. This second stage is, at present, an interactive software demonstration only: the student-facing application and AI tutor interface are functional, but the mesh propagation layer is simulated rather than field-validated, and no physical multi-device pilot has yet been conducted. This paper documents the design, method, and results of both stages, evaluates the technical trade-offs involved in the hardware transition, and proposes a framework for scaling seed-node-plus-mesh architectures to campus-wide and community-wide deployment.