♟️ MagnOSS – Your Open Source Chess Coach

Welcome to MagnOSS, our hackathon project built during {Tech: Europe}. Think of it as your very own Magnus Carlsen at home – free, open source, and powered by AI.

🌐 Waitlist Registration: https://magnoss.vercel.app/

🎥 Demo Youtube Video: https://magnoss.vercel.app/demo

---

🚀 What is MagnOSS?

MagnOSS is a web application designed to make learning chess fun, interactive, and accessible to everyone. Instead of boring books, overly complex websites, or expensive private lessons, MagnOSS helps you improve your chess skills through AI coaching, personalized analysis, and theory exploration.

---

✨ Features

🧑‍🤝‍🧑 Play & Learn

• Play against our custom-built chess AI.
• After each game, a virtual trainer highlights key mistakes and explains alternative moves.
• Thanks to text-to-speech, the trainer actually talks to you.

🔗 Chess.com Integration

• Connect your chess.com account.
• Import and analyze your past games with the trainer.
• Interact with the analysis through text or voice.
• Track your growth in a personal dashboard powered by D3.js.

📚 Theory Mode

• Explore chess concepts like openings, middlegame strategies, endgames, and tactics.
• Learn through interactive visuals: arrows, colored boxes, and step-by-step guidance.
• Ask questions by text or voice and get clear, AI-powered explanations.

---

🛠️ Under the Hood

MagnOSS combines traditional chess engines with cutting-edge AI:

• Stockfish → for algorithmic move evaluation.
• Custom RAG pipeline → over 1,600 documents scraped with Lightpandas + Puppeteer.
• Dust agent + OpenAI API → to classify, filter, and tag chess knowledge (openings, tactics, endgames, etc.).
• Custom-built UI → from the chessboard to the trainer interface, everything is hand-crafted.

---

📦 Prerequisites

Before running MagnOSS, make sure you have:

• Python 3.x

• Install dependencies:

  pip install -r requirements.txt

• Stockfish: download the binary from stockfishchess.org and set the correct path in backend/models/stockfish.py.

• A modern web browser (Chrome, Firefox, Edge, etc.).

• An .env file at backend/src/rag/.env containing:

  WEAVIATE_REST_ENDPOINT=your_weaviate_instance
  WEAVIATE_API_KEY=your_weaviate_api_key
  OPENAI_API_KEY=your_openai_api_key

---

▶️ Running the Project

1. Run backend

   cd backend
   python server.py

2. Run frontend

   cd frontend
   python -m http.server

3. Open in browser Go to http://localhost:8000/

---

🎯 Why MagnOSS?

We wanted to create a tool that:

• Feels like having a real coach by your side.
• Makes chess theory visual and fun.
• Is open source, so the community can improve and extend it.

---

🌍 Built at {Tech: Europe}

MagnOSS was designed and prototyped during the {Tech: Europe} hackathon. Every line of code, every design choice, and every integration was made in just a few intense days of collaboration.

---

💡 What’s Next?

• Expanding the theory base with even more annotated games.
• Mobile-first version for training on the go.
• Multiplayer “coach vs coach” mode for fun sparring with friends.

---

🏆 Credits

Developed by Shashwat Sharma, Enzo Pinchon, Stepan Svirin and Tristan Donzé for {Tech: Europe} hackathon. Special thanks to OpenAI, Dust, Stockfish, and the chess.com community for inspiration and tools.