DonGobbi/MwanaAI

MwanaAI — an AI-powered personal tutor for students in Malawi. Adapts to your class/form level and subject, explains step by step, reads photos of your homework, and generates practice questions. # MwanaAI — AI Learning Platform 🇲🇼 MwanaAI is an AI-powered learning platform for students in Malawi. It started as a personal tutor and has grown into a full learning loop — **Learn → Practice → Get help → Track progress** — with dashboards for **students, teachers and parents**, all aligned to the Malawi curriculum (Standard 1–8 / Form 1–4). The goal: make learning feel like a patient, encouraging human tutor — and give teachers and parents real tools to support each student. --- ## ✨ Features ### For students - **AI Tutor** — ask anything and get step-by-step help pitched at your class level. Supports: - 📷 **Homework photo upload** — snap a picture of your work and a vision model reads it and helps. - 🎤 **Voice** — speak your question and 🔊 listen to answers read aloud. - 💾 **Save & resume** chats. - Quick actions (*explain simply*, *another example*, *quiz me*, …). - **Guided lessons** — pick a subject and get a syllabus-ordered list of topics, each taught as a full lesson; mark topics complete and watch a progress bar. - **Adaptive quizzes & exam prep** — auto-marked multiple-choice quizzes by subject, with **PSLCE / JCE / MSCE** exam styles. An **Adaptive** mode tunes difficulty to your past scores, and **AI feedback** tells you exactly what to review after each quiz. - **Smart Study Plan** — AI reads your quiz performance and builds a personalised plan for what to study next. - **Progress & motivation** — scores by subject, 🔥 **day streaks**, a class **leaderboard**, and **achievement badges**. - **Assignments** — see tasks set by your teachers and complete them in one tap. ### For teachers - **Generate from your materials** — upload the syllabus, textbook pages, a screenshot, a PDF or Word doc, paste text, **or drop in a YouTube link**, and MwanaAI builds a lesson plan, quiz, revision notes or exam questions **grounded in that material** (not generic). Images are read by the vision model; PDFs/Word are parsed in-browser; YouTube transcripts are fetche …