SourcebookLM

An AI-powered interactive notebook application that helps you create, organize, and analyze content from various sources. SourcebookLM combines document processing, AI-generated insights, and interactive learning tools to transform your documents, URLs, and YouTube videos into structured knowledge.

Features

• Multi-Source Content Ingestion

  • Upload and extract text from PDF, DOCX, DOC, TXT, MD, CSV, and LOG files
  • Extract content from web URLs
  • Download and process YouTube video transcripts

• AI-Powered Tools

  • Chat: Interactive AI assistant with context awareness
  • Mind Maps: Generate hierarchical mind maps from your content
  • Reports: Create structured reports with executive summaries and findings
  • Flashcards: Generate educational flashcards for studying
  • Quiz: Create multiple-choice quizzes with explanations
  • Slides: Generate presentation slide decks

• Notebook Management

  • Create and organize multiple notebooks
  • Save and load notebook states
  • Persistent storage with MongoDB
  • Track sources and conversation history

Prerequisites

• Python 3.8 or higher
• MongoDB database (local or cloud-based like MongoDB Atlas)
• Google Gemini API key

Installation

1. Clone the repository

   git clone https://github.com/truepushkar/SourcebookLM.git
   cd SourcebookLM

2. Create a virtual environment (recommended)

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies

   pip install -r requirements.txt

4. Set up environment variables

   Create a .env file in the root directory based on sample.env:

   cp sample.env .env

   Edit .env and add your credentials:

   MONGO_URI=mongodb://localhost:27017/  # Or your MongoDB Atlas connection string
   GEMINI_API_KEY=your_gemini_api_key_here
   

Configuration

MongoDB Setup

• Local MongoDB: Install MongoDB locally and use mongodb://localhost:27017/
• MongoDB Atlas: Create a free cluster at mongodb.com/cloud/atlas and use the provided connection string

Gemini API Key

1. Visit Google AI Studio
2. Create a new API key
3. Add it to your .env file

Usage

1. Start the application

   python app.py

2. Access the application

   Open your web browser and navigate to:

   http://localhost:5000
   

3. Create a notebook

   • Click "New Notebook" from the dashboard
   • Add sources by uploading files, entering URLs, or YouTube video links
   • Use the AI tools to analyze and transform your content

4. Interact with your content

   • Chat with the AI about your sources
   • Generate mind maps to visualize concepts
   • Create reports and summaries
   • Build flashcards for study sessions
   • Generate quizzes to test your knowledge

Project Structure

SourcebookLM/
├── app.py                 # Main Flask application
├── generative.py          # AI/Gemini integration
├── upload_text.py         # Document and content extraction
├── requirements.txt       # Python dependencies
├── sample.env            # Environment variables template
├── templates/            # HTML templates
│   ├── dashboard.html    # Main dashboard
│   └── notebook.html     # Notebook interface
└── static/              # Static assets
    ├── styles.css       # Stylesheets
    ├── app.js          # Main application logic
    ├── dashboard.js    # Dashboard functionality
    ├── mindmap.js      # Mind map visualization
    ├── report.js       # Report generation
    ├── flashcards.js   # Flashcard interface
    └── quiz.js         # Quiz functionality

API Endpoints

Notebook Management

• GET /api/notebooks - List all notebooks
• POST /api/notebooks - Create or update a notebook
• GET /api/notebooks/<id> - Load a specific notebook
• DELETE /api/notebooks/<id> - Delete a notebook

Content Processing

• POST /api/upload - Upload and extract text from files
• POST /api/extract_url - Extract content from a URL
• POST /api/extract_youtube - Extract YouTube video transcript

AI Tools

• POST /api/chat - Chat with AI assistant
• POST /api/tool_action - Generate content with AI tools (mindmap, reports, flashcards, quiz, slides)

Technologies Used

• Backend: Flask (Python web framework)
• Database: MongoDB (document storage)
• AI: Google Gemini API (generative AI)
• Document Processing: PyMuPDF, python-docx, mammoth
• Web Scraping: BeautifulSoup4, requests
• YouTube: youtube-transcript-api
• Frontend: HTML, CSS, JavaScript

License

This project is licensed under the MIT License. See the LICENSE file for details.

Author

Pushkar Singh

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Support

If you encounter any issues or have questions, please open an issue on the GitHub repository.

Acknowledgments

• Google Gemini API for AI capabilities
• MongoDB for database solutions
• The open-source community for the excellent libraries used in this project