This paper presents a real-time, low-cost human–computer interaction system enabling hands-free cursor control through eye-gaze and blink detection. The system integrates MediaPipe FaceMesh for facial landmark tracking and OpenCV for live image processing, allowing accurate gaze estimation and multi-blink classification. Left, right, and double blinks are mapped to distinct mouse click operations, while gaze coordinates are translated into cursor movement using adaptive acceleration, deadzone filtering, and smoothing to reduce jitter. Experimental evaluation with ten participants demonstrated cursor accuracy between 80–93% and blink recognition rates exceeding 90%. Visual analyses confirmed system stability, responsiveness, and adaptability. Blink-to-click latency remained below 0.25 seconds, and no recalibration was required across sessions. Designed to run on consumer-grade hardware without specialized sensors or infrared illumination, the solution offers a calibration-free, hardware-independent alternative to traditional gaze-tracking systems. Its low computational cost and real-time performance make it suitable for assistive technologies, rehabilitation tools, and touch-free computing environments.