MicroTutor AI 📘

An intelligent learning system powered by Google Gemini File Search API that provides RAG-based tutoring, flashcards, quizzes, and study schedules from your PDF documents.

App Features

PDF Upload & Indexing

Upload and index PDF documents for processing. The interface shows real-time status updates during indexing and displays all indexed PDFs with their store IDs.

Visual Study Schedule Generator

Generate intelligent, multi-day study schedules from uploaded PDFs. Each day includes session titles, summaries, and page references to guide your learning.

Interactive Flashcards

Create and study with interactive flashcards. Navigate through cards, flip to see answers, and review key concepts from your PDFs.

🚀 Quick Start

Prerequisites

• Python 3.8+
• Virtual environment (venv)
• API Keys:
  • GEMINI_API_KEY (Required) - Get from Google AI Studio
  • OPENAI_API_KEY (Optional) - For mem0 memory features
  • MEM0_API_KEY (Optional) - For mem0 memory features

Setup (Windows PowerShell)

1. Activate your virtual environment:

   .\.venv\Scripts\Activate.ps1

2. Run the setup script:

   .\setup.ps1

   Or manually install dependencies:

   pip install -r requirements.txt

3. Configure API keys:

   Create a .env file in the project root:

   GEMINI_API_KEY=your_gemini_api_key_here
   OPENAI_API_KEY=your_openai_api_key_here  # Optional
   MEM0_API_KEY=your_mem0_api_key_here      # Optional

   Or use env_template.txt as a template.

4. Run the application:

   python app.py

5. Access the UI:

   • The Gradio interface will open in your browser automatically
   • Or visit the URL shown in the terminal (usually http://127.0.0.1:7860)

📋 Features

• 📄 PDF Upload & Indexing: Upload multiple PDFs and index them using Gemini File Search
• 🤖 RAG Tutor Chat: Ask questions and get answers with page citations
• 📅 Intelligent Study Schedule: Generate personalized study schedules from your PDFs
• 🃏 Flashcards: Generate interactive flashcards for any topic
• 📝 Interactive Quiz: Create and take quizzes with immediate feedback
• 🧠 Learning Memory: Track your progress with mem0

🏗️ Architecture

System Overview

MicroTutor
├── app.py                    # Entry point
├── core/
│   ├── file_search.py       # Gemini File Search integration
│   ├── page_map.py          # PDF page mapping
│   ├── tutor.py             # RAG tutor assistant
│   ├── schedule.py          # Study schedule generator
│   ├── flashcards.py        # Flashcard generator
│   ├── quiz.py              # Quiz generator
│   └── memory.py            # Memory manager (mem0)
└── ui/
    └── interface.py         # Gradio UI

Gemini File Search Integration

Gemini File Search API supports multiple input formats (JSON, JavaScript, Python) and processes PDF documents through advanced neural network models for intelligent content understanding and retrieval.

RAG Architecture (Indexing & Querying)

The RAG (Retrieval-Augmented Generation) architecture used by MicroTutor:

• Indexing Process (Offline): Documents are stored, embedded, and indexed in a vector database
• Querying Process (Real-time): Gemini intelligently decides when to retrieve context from the knowledge base, generates search queries, and synthesizes answers with citations

🔧 Troubleshooting

ModuleNotFoundError

If you get ModuleNotFoundError:

1. Ensure your virtual environment is activated (check for (.venv) in your prompt)
2. Run pip install -r requirements.txt inside the venv
3. Or run .\setup.ps1 to automate setup

API Key Errors

• Ensure .env file exists in the project root
• Check that GEMINI_API_KEY is set correctly
• Verify API keys are valid and have proper permissions

Import Errors

• Make sure you're using the Python from your venv:

  python -c "import sys; print(sys.executable)"

• Should show path to .venv\Scripts\python.exe

📝 Requirements

See requirements.txt for full dependency list:

• google-generativeai - Gemini API client
• gradio - Web UI framework
• pymupdf - PDF processing
• mem0ai - Memory/persistence (optional)
• python-dotenv - Environment variable management

📄 License

MIT License