Skip to Main Content

Hocam.net

Built at Meta Global AI Developer Hackathon · Oct 26, 2026 · Remote

Hocam.net — Demo video

Hocam.net is an AI-assisted exam preparation platform for Turkish and US exam systems. Students preparing for TYT, AYT, DGS, SAT and ACT often depend on expensive printed question banks, carry heavy backpacks, and practice with one-size-fits-all content. Families typically spend TRY 5,000–7,000 per year on printed preparation: about TRY 2,500–3,000 for 4–6 subject question banks, TRY 1,500–2,500 for topic test books, and TRY 1,000–1,500 for printed mock exams and worksheets. Students still lack targeted feedback, and teachers lack a clear view of gaps. Hocam.net replaces this with a digital study system. It generates curriculum-aligned questions, verifies each question with an independent second solution, adapts difficulty to the student’s level, schedules spaced review, and builds a weekly study plan. A mistake notebook brings incorrect questions back at 1, 3, 7, 14 and 30-day intervals. A diagnostic test gives an initial view of strengths and gaps. The AI Coach helps students decide what to study today, while topic summaries and step-by-step solutions explain the reasoning behind each answer. Teachers can create classes, share a class code, follow student progress and use a class-level XP leaderboard. The environmental cost is also significant. Producing one ton of paper requires roughly 15–20 trees; Türkiye consumes about 6 million tons of paper and cardboard annually, and global paper production is linked to billions of trees harvested each year. Even a partial shift to digital practice can avoid large volumes of paper and printing, reducing pressure on forests and natural habitats. The prototype is built with PHP 8.1 and MySQL, with JSON-based content storage. It uses Meta’s Model API (muse-spark-1.3) for question generation, independent answer verification, explanations, summaries and AI Coach conversations. Structured JSON prompts make generated content reliable to parse; a second solving pass checks whether the answer key agrees. Token usage is logged per request. The interface supports Turkish and English, responsive mobile/tablet layouts, speech output, screen-reader-friendly labels and keyboard navigation. The codebase includes 3,000+ automated checks and a working teacher/student class flow. Current stage: working prototype with a live demo. The next step is a real classroom pilot to measure learning gains, retention and teacher workflow. The long-term goal is to reduce annual personalized learning costs from roughly TRY 5,000–7,000 to around TRY 100 per student and to extend the same adaptive architecture beyond exam preparation into broader school subjects and lifelong learning. Accessibility: Hocam.net includes speech output that reads questions and solutions aloud, screen-reader-friendly labels and keyboard navigation, so visually impaired students can study independently. Live demo: https://www.hocam.net/meta/

Team