SmartStudy AI 🧠✨

An intelligent multi-modal educational platform powered by Google Generative AI (Gemini) and built with Python & Streamlit. It transforms complex study materials (PDF documents, slides, images, and raw text) into comprehensive narrative summaries, automated MCQ quizzes with detailed explanations, and active-recall flashcards.

Python Version Framework AI Model Architecture License

• 📖 Narrative Multi-modal Synthesis: Ingests PDFs, images, and text to generate structured summaries highlighting Exam Hotspots and intuitive analogies.
• 📝 Automated Quiz & Exam Simulation: Dynamically creates 10-question MCQ quizzes across customizable difficulty levels with instant grading and referenced rationales.
• ⚡ Active Recall Flashcards: Automatically extracts technical terms and definitions into bite-sized revision flashcards.
• 🛡️ Production-Grade Resilience:
  • Exponential Backoff: Configured with tenacity retry logic to gracefully handle rate limits (429) and transient network errors.
  • Strict JSON Schema Enforcement: Robust regex parsing and validation ensure 100% structured JSON integrity without UI crash.
• 🎨 Glassmorphism Dark UI: Clean, modern responsive interface crafted with custom CSS styling.

graph TD
 A[Input: PDF / Image / Text] --> B[SmartStudy Engine]
 B --> C{Task Type}
 C -->|Summary| D[Narrative Prompt + Hotspots]
 C -->|Quiz| E[MCQ Prompt + Strict JSON Schema]
 C -->|Flashcards| F[Concept Extraction Prompt]
 D --> G[Gemini API with Tenacity Retry]
 E --> G
 F --> G
 G --> H[JSON Schema Validation & Fallback]
 H --> I[Streamlit Glassmorphism UI]

git clone https://github.com/Kong19565/SmartStudy_AI.git
cd SmartStudy_AI

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

pip install -r requirements.txt

Create a .env file in the project root:

GEMINI_API_KEY="your_actual_gemini_api_key"

streamlit run app.py

• Core Engine: Python 3.10+, Google Generative AI SDK (google-generativeai)
• Frontend / Dashboard: Streamlit, Custom Glassmorphism CSS
• Resilience & Validation: Tenacity (Retry & Backoff), Pydantic / JSON schema parsers
• Document Processing: PyPDF2, Pillow (PIL)

Apisak Kongphakdee (Kong)

• GitHub: @Kong19565
• Location: Bangkok, Thailand