The Nepali Sign Language Recognition System is a real-time machine learning application developed to reduce the communication gap between speaking and non-speaking or non-hearing communities. Unlike many existing systems that focus on globally recognized sign languages, this project specifically targets Nepali Sign Language. Using MediaPipe for hand landmark detection and a Feedforward Neural Network (FNN) model for gesture classification, the system accurately recognizes hand gestures and converts detected text into speech using Google’s Text-to-Speech API. The model achieved a validation accuracy of 97.43% across 60 gesture classes, demonstrating reliable performance. This system has practical applications in teaching, learning, and promoting inclusivity for the deaf and non-speaking community in Nepal.