This paper describes the design and implementation of an affordable vision-based Augmentative and Alternative Communication (AAC) system for paralysis patients who retain eye gaze or blinking control to communicate. The proposed system employs a smartphone camera and a regular computer to detect and analyze eye movements using the Haar Cascade algorithm for real-time facial feature detection. These eye movements are translated into text and then converted into speech using a text-to-speech (TTS) mechanism. Four key eye movements (blink, left, right, up) are utilized to develop communication blocks, enabling individuals with minimal muscle control to communicate easily. Experimental results with 12 participants across varying educational and technological literacy levels demonstrate the system's ability to provide efficient and accurate communication, with participants showing varying degrees of success based on their familiarity with digital tools. The findings highlight the system's potential to enhance communication for individuals with severe motor impairments and emphasize the need for further research into adaptive interfaces and personalized calibration to improve accessibility.