Back to Projects

StreamlyAI

Converts YouTube videos into searchable, conversational content via transcription, semantic chunking, RAG-based Q&A, and auto-generated highlight reels.

StreamlyAI screenshot 1

Project Overview

StreamlyAI is a full-stack generative AI platform that turns YouTube videos into searchable, interactive content. Smart transcription runs via AssemblyAI, topic segmentation and semantic search use ChromaDB, and a conversational Q&A assistant is powered by RAG pipelines and LLMs (Groq, Cohere). Features include auto highlight-reel generation, real-time querying over the transcript, and speaker-based segmentation. Built with React + TypeScript on the frontend and Flask on the backend, with production-ready APIs.

Key Features

  • Real-time transcript analysis
  • Video summarization and topic extraction
  • Sentiment analysis
  • Chatbot for video Q&A
  • Reel and short creation
  • Speech-to-text with Whisper/AssemblyAI
  • Vercel and Azure deployment

Case Study

Problem

Content creators and researchers need efficient tools to analyze, summarize, and interact with long-form video and podcast content. Manual review is slow and often misses key insights.

Solution

StreamlyAI automates transcript generation, topic extraction, and summarization using state-of-the-art NLP models. It also provides a chatbot interface for real-time Q&A about the content and tools for generating short-form video highlights.

Research & Techniques

  • Speech-to-Text: Used Whisper and AssemblyAI for accurate, real-time transcription.
  • Transformer Models: Leveraged HuggingFace Transformers and Cohere embeddings for topic modeling, sentiment analysis, and summarization.
  • RAG (Retrieval-Augmented Generation): Combined transcript retrieval with LLMs for context-aware Q&A.
  • Video Processing: Used pytube for video downloads and custom scripts for highlight generation.

    • Results

  • Enabled content creators to generate summaries and highlights 10x faster.
  • Improved accessibility for podcast and video content.
  • Demonstrated real-world application of advanced NLP and RAG techniques in academic settings.

    • Related Projects

    Smart Hire AI

    Smart Hire AI

    AI-powered recruitment platform with active enterprise pilots. Best FYP Award — UMT Software Engineering Department 2026.

    Vibe Terminal

    Vibe Terminal

    Cross-platform AI terminal assistant — execute shell commands from natural language, with voice I/O and multi-step agent workflows.

    Smart Grade AI

    Smart Grade AI

    AI-powered LMS with RAG-based mentoring — Microsoft Imagine Cup 2025 Global Semifinalist, backed by $25K Azure credits from Microsoft for Startups, in institutional rollout at UMT.