YouTube Video Digest AI

Welcome to the YouTube Video Digest AI repository! This project allows users to digest YouTube videos using Google Gemini or local LLM models via Ollama, and interactively ask questions about the content of the provided video URL.

Features

• Digest YouTube Videos: Analyze video content through Google Gemini or a local LLM (Ollama).
• Interactive Q&A: Ask questions related to the video and receive informative answers.
• Local LLM Support: Use a locally hosted LLM (Ollama) with various pre-loaded models for context-based interaction.
• Download MP4/MP3: Download Youtube video as MP4 or MP3

Getting Started

Prerequisites

Before you begin, ensure you have the following installed on your machine:

• Node.js (v18 or higher)
• Yarn
• yt-dlp
• Ollama (for local LLM support)

Installation

1. Clone the repository:

   git clone https://github.com/zahnno/youtube-helper-ai
   cd youtube-helper-ai

2. Setup the Environment Variables (GEMINI):

   • Create a .env file in the server directory.
   • Add your Google Gemini API key:

     GEMINI_API_KEY=your_gemini_api_key
     

3. Install Dependencies:

   • Navigate to the server folder and install dependencies:

     cd server
     npm install

   • Navigate to the web-client folder and install dependencies:

     cd ../web-client
     npm install

4. Install yt-dlp: For detailed installation instructions, visit the yt-dlp GitHub page.

5. Run Ollama for Local LLM:

   • Make sure Ollama is installed on your machine.
   • Start Ollama locally on localhost:11434:

     ollama serve

Running the Application

1. Start the Server:

   • Navigate back to the server folder and run:

     cd ../server
     npm run dev

2. Start the Web Client:

   • In a new terminal window, navigate to the web-client folder and run:

     cd ../web-client
     npm run start

Your application should now be running locally! Open your browser and go to http://localhost:3000 to interact with the YouTube Video Digest AI. You can select between Google Gemini or a local LLM model (via Ollama) for chatting in the context of the video.

Local LLM Models

When Ollama is running, you can select from a list of loaded models to interact with. Make sure the models are properly loaded into Ollama before starting the application.

License

This project is licensed under the MIT License - see the LICENSE file for details. Contributions are welcome!

Acknowledgments

• Google Gemini
• Ollama
• yt-dlp