<p>This paper presents a method for generating speaking avatars from text and images, designed for advanced virtual interaction. The process consists of two key components: generating audio from text and generating video from an image and the generated audio. Facial recognition and animation algorithms are applied to transform a static image into a dynamic, lifelike avatar capable of realistic mouth movement and expressions, with the avatar’s lip movement precisely synchronized with the generated audio, aimed at creating a seamless and engaging user experience. The analysis of performance of video generation using advanced hardware configurations highlights the importance of GPU acceleration. Tests with varying batch sizes and quality settings reveal the challenges in efficient processing, making the characteristics of tools for video creation and audio synthesis a key element in producing high-quality multimedia content. By leveraging powerful CPUs and GPUs, the study demonstrates how current technology enables efficient and sophisticated video and audio generation, underscoring the need for optimized resource management to achieve balanced processing speed and output quality.</p>