---

🧠 Axiom Math AI

The Adaptive, Pedagogy-First Math Tutoring Engine

Axiom Math AI is a full-stack, LLM-powered platform that bridges the gap between raw AI and structured education. It features an adaptive tutoring engine that adjusts its cognitive load based on student needs, all wrapped in a secure, high-performance architecture.

---

🚀 Core Features

🎓 Adaptive Tutoring Logic

• Cognitive Slicing: Dynamically manages LLM context windows based on difficulty.

• Easy: Deep history for step-by-step hand-holding.

• Hard: Minimal history to challenge student logic.

• KaTeX Integration: Native rendering of high-fidelity LaTeX for Calculus, Trigonometry, and Algebra.

• IQ Warmup: A pre-login math challenge that sets the academic tone.

🛡️ Production-Grade Security

• PII Scrubbing: Automated regex filters to protect student privacy before data hits the LLM.
• Domain Guardrails: Proprietary logic that restricts the AI to mathematics, preventing off-topic drift.
• JWT Auth: Secure, stateless session management using HS256 symmetric encryption.

---

🛠️ Tech Stack

• Frontend: React (Vite), Axios, KaTeX, CSS3 (Glassmorphism UI).
• Backend: FastAPI (Python), Uvicorn, SQLAlchemy.
• AI Engine: OpenAI GPT-4o-mini (via OpenRouter).
• Deployment: Render (API) & Vercel (Frontend).

---

💻 Local Setup & Installation

Follow these steps to clone and run the project on your local machine.

1. Clone the Repository

git clone https://github.com/YOUR_USERNAME/axiom-math-ai.git
cd axiom-math-ai

2. Backend Setup (FastAPI)

# Navigate to backend
cd math-tutor-backend

# Create and activate virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Create .env file
touch .env

Add the following to your .env:

OPENROUTER_API_KEY=your_key_here
JWT_SECRET_KEY=your_secret_string
ALGORITHM=HS256

3. Frontend Setup (React/Vite)

# Open a new terminal in the root folder
cd math-tutor-frontend

# Install dependencies
npm install

# Run the development server
npm run dev

---

🏗️ Architecture Flow

1. Request: Client sends a JWT-authenticated JSON payload.
2. Safety: Backend validates math-relevance and scrubs PII.
3. Inference: Sliced context is sent to the LLM for step-by-step problem solving.
4. Response: Frontend parses the response into interactive step-cards and rendered math blocks.

---

📝 Roadmap

• Vector Memory: Implementing a Vector DB (Pinecone) for long-term student progress tracking.
• Voice Mode: Real-time speech-to-math interaction.
• Teacher Dashboard: Analytical view of student struggle-points.

---

📄 License

Distributed under the MIT License. See LICENSE for more information.

---