AIthena is an AI-powered peer-assisted learning platform that identifies conceptual gaps in student understanding and delivers personalized micro-lessons, quizzes, and visual aids. It combines generative AI, speech recognition, and gamified learning to enhance how students study and retain concepts.
- Upload notes, past year questions (PYQs), or textbooks
- LLM (Mistral) generates concept-check questions from the material
- Users answer via text or speech
- LLM analyzes the response, identifies weak areas, and delivers targeted micro-lessons
- Upload handwritten or typed notes
- Text is extracted using OCR (
pytesseract) - LLM converts the extracted content into a personalized quiz
- Users answer quizzes via microphone input
- Speech is transcribed using
whisper.cpp - LLM compares the response to a reference answer and identifies conceptual gaps
- Upload syllabus and PYQs
- Extracts and matches topics using TF-IDF and semantic similarity
- Generates topic importance metrics, visualized using bubble charts
- Takes into account available time, exam date, and weak topics
- Outputs a customized, day-wise study plan
- Generates revision notes and mindmaps for each topic
| Component | Technology |
|---|---|
| Backend | Flask (deployed on Render) |
| Frontend | HTML, CSS, JavaScript (GitHub Pages / Netlify) |
| LLM | Ollama (Mistral, Phi, Gemma) |
| Speech-to-Text | whisper.cpp (Tiny model) |
| OCR | pytesseract |
| Data Visualization | Chart.js, D3.js |
| UI Design | Figma |
# Clone the repository
git clone https://github.com/AakxSha/AIthena.git
cd AIthena/backend
# Set up virtual environment
python -m venv venv
source venv/bin/activate # Use venv\Scripts\activate on Windows
# Install dependencies
pip install -r requirements.txt
# Run the Flask app
python app.py