Paper Title
AI-POWERED YOUTUBE VIDEO ANALYSIS SYSTEM FOR TITLE AND DESCRIPTION OPTIMIZATION

Abstract
This paper presents the YouTube Video Analysis System, a web-based application developed to enhance video titles, descriptions, and engagement tactics for content creators. Utilizing an intuitive interface developed with HTML, CSS, and JavaScript, users can seamlessly upload videos, while the backend, powered by Flask, manages the processing workflow. The system extracts audio using FF mpeg and converts speech to text through the Speech Recognition library in conjunction with Google’s Web Speech API. By employing Distil BERT for keyword extraction, the application generates SEO-optimized titles and descriptions, while also predicting engagement metrics via the YouTube Data API. The final results are displayed in a user-friendly format, enabling creators to enhance their visibility and audience engagement on the platform. Keywords - YouTube Video Analysis System, Video Optimization, SEO (Search Engine Optimization), DistilBERT, Keyword Extraction, Engagement Prediction.