TalkTube
Ahsan Umar
🎥 YouTube Content Assistant
A powerful Streamlit application that allows users to analyze and interact with YouTube video content through natural language questions.
🌟 Features
YouTube Video Processing: Input any valid YouTube URL to analyze its content
Audio Transcription: Automatically transcribes video content using Whisper AI
Interactive Q&A: Ask questions about the video content using advanced RAG (Retrieval-Augmented Generation)
GPU Acceleration: Utilizes CUDA for faster processing when available
User-Friendly Interface: Clean and intuitive Streamlit interface
🚀 Installation
Clone the repository:
Install the required dependencies:
Set up environment variables: Create a
.env file in the root directory with necessary API keys and configurations.📦 Dependencies
streamlit
streamlit-extras
yt-dlp
whisper
torch
python-dotenv
(other dependencies as specified in requirements.txt)
🎮 Usage
Run the Streamlit app:
Enter a YouTube URL in the input field
Wait for the video to be processed and transcribed
Ask questions about the video content in the chat interface
💡 How It Works
Video Processing: The app downloads the audio from YouTube videos using yt-dlp
Transcription: Uses OpenAI's Whisper model to transcribe the audio content
Question Answering: Implements RAG (Retrieval-Augmented Generation) to provide accurate answers based on the video content
🛠️ Technical Details
Built with Streamlit for the web interface
Uses Whisper AI for accurate speech-to-text transcription
Implements advanced RAG techniques for question answering
Supports both CPU and GPU processing
Handles various YouTube URL formats
⚠️ Notes
Processing time depends on video length and available computing resources
GPU acceleration significantly improves transcription speed
Internet connection required for YouTube video download and processing
Like this project
Posted Feb 4, 2025
A powerful Streamlit application that allows users to analyze and interact with YouTube video content through natural language questions. - codewithdark-git/Ta…
