AI READMATE

About the Project

In primary education, individual reading assessment remains a resource-intensive challenge, particularly in multilingual classrooms where high student-to-teacher ratios limit one-on-one instructional time. To address this gap, AI READMATE is developed as a web-based, real working prototype designed to function as both an interactive reading partner for young learners and an automated diagnostic co-pilot for educators. Supporting English, Bahasa Melayu, and Tamil across four progressive difficulty tiers—ranging from single words and sentences to paragraphs and essays—the system integrates live speech recognition with comparative AI analysis. When a pupil reads aloud into a standard device, AI READMATE instantly transcribes the speech, contrasts it against the canonical text, and detects reading hurdles such as missing articles, skipped words, and pronunciation errors. Moving beyond a conventional scoring mechanism, the system generates a unique Learning Fingerprint that diagnoses specific linguistic challenges and immediately translates them into targeted 3-minute micro-practice activities for real-time remediation. Furthermore, the centralized Teacher Dashboard automates initial classroom screenings by synthesizing pupil accuracy metrics, difficulty trends, and automated intervention recommendations into an intuitive interface. By combining low-hardware deployment requirements, multilingual inclusivity, and data-driven personalization, AI READMATE effectively reduces educator workloads while accelerating foundational literacy and fluency outcomes in primary schools.