🎥 VideoMath Tutor

Solve math problems from educational videos using OCR + AI!

VideoMath Tutor is a smart Chrome Extension + Python backend that captures paused video frames, extracts math expressions using OCR (Pix2Tex or Tesseract), and solves them using Together AI for detailed step-by-step solutions.

---

🚀 Features

• ⏸ Pause any video to auto-capture math problems
• 🔍 OCR via Pix2Tex CLI (preferred) or Tesseract fallback
• 🧠 AI-powered solving using Together AI (Mixtral / LLaMA)
• 🧾 Render LaTeX in an elegant KaTeX popup
• ✂ Copy, 🌐 Search, and ✅ Solve directly from overlay
• 💡 Built-in hint engine for learning context
• 🧲 Toggle extension on/off anytime

---

⚙ How It Works

1. User pauses a video
2. Content script captures the video frame
3. Frame sent to FastAPI backend at /ocr/single
4. OCR returns LaTeX (Pix2Tex or Tesseract)
5. User sees formatted math + options (solve, copy, hints)
6. On clicking Solve, /solve sends it to Together AI
7. Solution is cleaned and shown inline 🎯

---

🛠 Installation

Backend (FastAPI + Together AI)

# Clone the repo
git clone https://github.com/jayjain4554/VideoMath-Tutor.git
cd VideoMath-Tutor/backend

Create virtual environment

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

Install dependencies

pip install -r requirements.txt

Run server

uvicorn main:app --reload

🧠 Note:

• Ensure you have Tesseract installed and in PATH
• Install Pix2Tex CLI for better OCR accuracy
• Add your Together AI API key to the environment (TOGETHER_API_KEY)

Server runs at http://127.0.0.1:8000

---

2️⃣ Chrome Extension

1. Go to chrome://extensions
2. Enable Developer Mode
3. Click Load Unpacked
4. Select the extension/ folder from this repo

📌 Then click the extension icon and Activate Extension from popup

---

📂 Project Structure

VideoMath-Tutor/
├── backend/
│   ├── main.py              # FastAPI app with OCR + Together AI solve
│   ├── requirements.txt     # Python dependencies
│   ├── .env                 # TOGETHER_API_KEY (secure)
│   └── ocr_engine/          # Optional: Pix2Tex local installation
│
├── extension/
│   ├── content.js           # Core script to capture & display results
│   ├── manifest.json        # Chrome extension setup
│   ├── popup.html/.js/.css  # UI to toggle extension
│   └── katex.min.js/.css    # For math rendering

---

📸 Screenshots

---

🧠 AI Solving Engine

The backend uses the Together AI Inference API to solve LaTeX equations via models like:

• 🔸 mistralai/Mixtral-8x7B-Instruct-v0.1 (default)
• ✳️ Easy to upgrade to LLaMA 3 or GPT-NeoX

It returns clean, step-by-step solutions that are parsed into human-readable output using a custom LaTeX cleaner.

---

💡 Future Roadmap

• 📢 Add LaTeX-to-speech for accessibility
• 🤖 Plug into Wolfram Alpha API for verified math
• 📈 Build a learning dashboard and insights tracker
• 🔍 Interactive step-by-step explanation viewer

---

👨‍💻 Author

Made with ❤️ by Jay Jain Feel free to contribute, suggest, or reach out for improvements.

Learn math from videos — smarter than ever. 🧠🎬➕