Dragoman AI: Real-Time Speech Translation for Educational
K Sika, J - Chandru, G - Madhappan, M. Sai Kumar Reddy M.Vikram · International Journal on Science and Technology · 2025
This paper presents Dragoman AI, a real-time speech-to-speech translation system designed to enhance educational accessibility for non-English-speaking students in India. The system integrates a Raspberry Pi microcontroller, Google Speech-to-Text, Google Translate, and gTTS APIs to capture spoken input in regional languages (e.g., Tamil, Telugu, Hindi), translate it into English, and synthesize audible English output. A 1.8-inch ST7735 TFT display provides visual feedback of translated text, improving comprehension in classroom settings. The hardware setup includes a USB microphone, earphones, and a portable casing, ensuring affordability and ease of use. By addressing language barriers, Dragoman AI promotes inclusivity, enabling students to participate actively in English-medium instruction. The paper discusses the system’s architecture, implementation, performance evaluation, and potential impact on educational equity, with future enhancements focusing on offline functionality and additional language support.