AI YOUTUBE VIDEO SUMMARIZER FOR INTELLIGENT KNOWLEDGE EXTRACTION FROM VIDEO CONTENT
Masik Ahamed, M and Sangeetha Radhakrishnan, R (2026) AI YOUTUBE VIDEO SUMMARIZER FOR INTELLIGENT KNOWLEDGE EXTRACTION FROM VIDEO CONTENT. AI YOUTUBE VIDEO SUMMARIZER FOR INTELLIGENT KNOWLEDGE EXTRACTION FROM VIDEO CONTENT, 10 (5). pp. 12-17. ISSN 2456-9348
paper9.pdf
Download (311kB)
Abstract
The AI YouTube Video Summarizer is a full-stack, AI-powered web application designed to intelligently process,
analyze, and summarize YouTube video content using Google Gemini large language models. The system supports
a dual-architecture deployment — a Streamlit-based Python frontend for rapid use and a Flask REST API paired
with a React + Vite frontend for production environments. Six distinct output modes are provided: Short Summary,
Detailed Summary, Full Explanation, AI-generated timestamped chapters, full transcript download, and a
Retrieval-Augmented Generation (RAG) chatbot for interactive Q&A over video content. The RAG pipeline uses
ChromaDB for in-memory vector storage and Gemini embeddings for semantic similarity search, enabling
factually grounded, hallucination-free answers. A mind map generation feature produces hierarchical Mermaid.js
diagrams for visual knowledge representation. User acceptance testing with 10 participants confirmed an average
satisfaction score of 4.6/5 across all features, with summarization latency under 8 seconds on the gemini-flash�latest model. The system demonstrates how open-source, zero-infrastructure-cost LLM tooling can deliver
enterprise-grade video intelligence at no subscription cost.
| Item Type: | Article |
|---|---|
| Subjects: | Computer Applications > Computer Science |
| Domains: | Computer Applications |
| Depositing User: | Mr IR Admin |
| Last Modified: | 16 May 2026 10:11 |
| URI: | https://ir.vistas.ac.in/id/eprint/19802 |
