ZENI

In Ethiopia, the barrier to digital education is rarely a lack of content; it is a lack of context. A student in Addis Ababa or Adama looking to learn "Introduction to React" or "High School Physics" on YouTube is immediately confronted with two adversaries: the algorithm and the infrastructure.

The algorithm is designed for engagement, not education. A search for a physics lecture often leads to a rabbit hole of entertainment, fragmenting focus. The infrastructure, specifically bandwidth reliability and data costs, punishes exploration. Every wrong click or low-quality video loads a cost on the student’s limited data plan.

Our research highlighted a critical gap: while 80% of urban students had smartphone access, their "time-on-task" for learning was shockingly low. They spent more time searching for quality content than consuming it. Teachers, meanwhile, struggled to assign digital resources because they couldn't track who watched what.

"

The internet is a library with all the books on the floor. We just need to put them on shelves.

— Ethiopian High School Teacher, during ZENI user interviews

The solution was not to host video content (which would be prohibitively expensive and redundant) but to index it. ZENI acts as a semantic layer on top of YouTube.

// Service/YouTubeImporter.php

public function convertPlaylistToCourse(string $playlistId, int $curatorId): Course
{
    $youtubeData = Laravel\Facades\Http::get("https://www.googleapis.com/youtube/v3/playlistItems", [
        'part' => 'snippet,contentDetails',
        'playlistId' => $playlistId,
        'key' => config('services.youtube.key'),
    ])->json();

    return DB::transaction(function () use ($youtubeData, $curatorId) {
        $course = Course::create([
            'title' => $youtubeData['items'][0]['snippet']['title'] ?? 'New Course',
            'curator_id' => $curatorId,
            'status' => 'draft',
        ]);

        $modules = collect($youtubeData['items'])->map(function ($item, $index) use ($course) {
            return new Module([
                'course_id' => $course->id,
                'title' => $item['snippet']['title'],
                'video_id' => $item['contentDetails']['videoId'],
                'order' => $index + 1,
                'duration_seconds' => 0, // Fetched via separate job to save API quota
            ]);
        });

        $course->modules()->saveMany($modules);
        
        return $course;
    });
}

Implementing ZENI required a stack that was modern yet resilient. We chose Laravel 11 for its robustness and Livewire 3 to create a "Single Page Application" (SPA) feel without the heavy JavaScript payload of a React/Vue app.

1. Solving for Bandwidth (The "Volt" Advantage): We utilized Livewire 3’s "Volt" functional API to keep component overhead minimal. Instead of sending large JSON payloads, the server renders the HTML and sends small diffs. This significantly reduces the data usage for students on 3G/4G networks. We also implemented aggressive caching for YouTube metadata, so we only hit the YouTube API once per course creation, not on every page load.

2. Real-Time Interaction with Reverb: To make remote learning feel communal, we used Laravel Reverb (WebSocket server). When a student finishes a module, their progress updates instantly on their dashboard without a page reload. More importantly, it enables "Live Sessions"—where a teacher can pause a video for all students simultaneously to explain a concept, synchronized via WebSockets.

3. The Ethiopian Fintech Layer: Monetization for premium courses was handled via Chapa. We integrated the Chapa payment gateway to allow students to pay via Telebirr or local bank cards. The integration was wrapped in a custom PaymentProvider service that handles webhooks securely, ensuring students get instant access upon payment confirmation—critical for trust in a developing digital market.

4. Admin Efficiency with Filament: We didn't build an admin panel from scratch. We used FilamentPHP, which saved us roughly 200 hours of development time. This allowed us to focus on the unique student-facing features rather than CRUD operations for user management.

In the beta launch with 50 high school students in Addis Ababa, the results validated our hypothesis. Average session duration increased by 300% compared to native YouTube usage. Students reported that the "distraction-free" interface felt like a "digital classroom" rather than a social media feed.