Orbit

Camera gesture based 1v1 skill games with crypto wagers, voice chat, and winner NFTs

Orbit

Created At

ETHOnline 2025

Project Description

Orbit is a revolutionary Web3 gaming platform that combines AI-powered gesture recognition with blockchain technology to create immersive 1v1 skill-based games. Players compete in real-time using their camera for gesture detection in games like Pushup Battle, Rock Paper Scissors, and Table Tennis, with crypto wagers and instant NFT rewards for winners. The platform features: AI-Powered Games: MediaPipe hand tracking and pose detection for gesture-based gameplay Real-time Multiplayer: WebRTC video chat and Socket.io synchronization Crypto Integration: Solidity smart contracts for escrow and Yellow Network for gasless transactions Winner NFTs: Automatic ERC721 certificate minting for game winners AI Referee: Fetch.ai uAgents for automated game result verification Professional UI: Modern design with 3-column layout (Your Video | Game Screen | Opponent Video) Players can create rooms, invite friends, choose game modes, and compete for crypto prizes while earning unique NFT certificates that prove their victories.

How it's Made

Frontend Stack: Next.js 16 with TypeScript for the main application Tailwind CSS for modern, responsive styling MediaPipe for AI-powered hand tracking and pose detection WebRTC for peer-to-peer video/audio communication Socket.io for real-time multiplayer synchronization Ethers.js v6 for blockchain integration Smart Contracts: Solidity contracts deployed on Base Sepolia Hardhat 3 for development and deployment OpenZeppelin for security standards (ReentrancyGuard, Ownable, ERC721) GameEscrow.sol: Handles wager escrow and winner payouts OrbitWinnerNFT.sol: ERC721 contract for minting winner certificates AI & Computer Vision: MediaPipe Hands: Real-time hand gesture detection for Rock Paper Scissors and Table Tennis MediaPipe Pose: Full-body pose detection for Pushup Battle counting TensorFlow.js: Object detection capabilities Custom gesture recognition: Hand-to-paddle mapping for Table Tennis Blockchain Integration: Yellow Network: Off-chain state channels for gasless transactions (simulated) Blockscout: Transaction explorer integration Base Sepolia: EVM-compatible testnet for deployment AI Referee Agent: Fetch.ai uAgents for automated game result verification Architecture Highlights: 3-Column Game Layout: Your Video | Game Screen | Opponent Video for immersive experience Real-time Synchronization: Socket.io events for score updates, game state, and chat Responsive Design: Works on desktop and mobile with camera access Professional UI: Dark theme with cyan/blue gradients, mouse-following effects Hackathon Integrations: Hardhat 3: Upgraded to latest version for prize eligibility Blockscout: SDK integration for transaction exploration ASI Alliance: Fetch.ai uAgents for AI referee system Yellow Network: Session-based gasless transaction simulation Notable Technical Achievements: Multi-game AI Integration: Seamlessly switches between different AI models (hands, pose, objects) Real-time Physics: Custom game engines for Table Tennis with collision detection NFT Metadata: Rich metadata including opponent, game mode, wager amount, and timestamp Cross-platform Compatibility: Works with MetaMask, Phantom, and other Web3 wallets The project demonstrates the future of Web3 gaming where skill, AI, and blockchain technology converge to create engaging, fair, and rewarding competitive experiences.

background image mobile

Join the mailing list

Get the latest news and updates