This paper introduces an enhanced multifactor authenticator built using face and voice biometrics. The system leverages advanced deep learning models such as SSD MobileNet V1 for facial detection and 68-point face landmark detection for precise facial recognition, achieving an accuracy of 98.5%. Users are required to verbally authenticate by reading a randomly generated phrase, incorporating voice biometrics with a verification accuracy of 97.8%. Liveness detection is achieved through the 68-point face landmark model to prevent spoofing attempts, ensuring the authenticity of user interactions. This dual-factor approach helps remove vulnerabilities in traditional authentication methods, thereby enhancing security and user convenience. Empirical results demonstrate the system’s overall accuracy of 99.5%, with an average authentication response time of 500 ms. This project significantly contributes to the field by integrating advanced biometric techniques to improve system security.