AI-Powered OCR-Based Translator with Speech Output for Partial Visual Impaired Users

In multilingual and diverse countries like India, individuals with visual impairments or low literacy levels face significant barriers in accessing printed text, especially in regional languages. Current screen readers and assistive technologies are largely limited to English, leaving inadequate support for regional languages such as Hindi, Kannada, Tamil, and Telugu. This lack of accessibility affects an estimated 4.95 million blind persons and over 70 million people with vision impairment in India, as well as more than 137 million with near vision loss, restricting their independence and participation in education, employment, and daily life. To address this challenge, we propose an AI-powered web application that converts text from images into audible speech in Indian regional languages, thereby bridging the accessibility gap. The solution integrates Tesseract.js for client-side optical character recognition (OCR), Google Translate API for translation, and the Web Speech API for text-to-speech synthesis. Optimized for mobile devices and low-bandwidth environments, the work proposed focuses on emphasizing speed, privacy, and ease of use. Simulated evaluations show OCR accuracy between 75–85% for clean printed text and 65–75% for moderately noisy images, with speech output latency under 500 ms per sentence, enabling near real-time interaction and offering practical support for millions of users across India