In today's digital landscape, video content is abundant across platforms, making manual viewing and analysis time-consuming and inefficient. This project presents an AI-powered Video Summarization System that leverages Spring Boot for scalable backend services and the Gemini API (Google's multimodal AI model) for generating intelligent, human-like video summaries. The system processes uploaded videos to extract keyframes and transcribes audio into text using advanced speech-to-text algorithms. These textual cues, along with frame metadata, are sent to the Gemini API, which analyses the context and generates concise, meaningful summaries using state-of-the-art natural language understanding. The backend, built using Spring Boot, handles video upload, storage, frame extraction, API communication, and summary delivery through RESTful endpoints. The architecture ensures modularity, scalability, and easy integration with frontend platforms or third-party applications. The summarization results include both text-based summaries and optionally highlight reel generation, improving accessibility and content consumption across domains such as e-learning, media, surveillance, and digital journalism. By combining traditional backend engineering with cutting-edge generative AI, the project demonstrates a powerful, automated pipeline for understanding and summarizing video content with minimal human intervention.

Keywords: Video Summarization, Gemini API, Content Compression, Video Analytics