Whal-E

Problem Statement

Whal-E is an AI-powered DeFi application focused on whale tracking and automated copy trading. Its purpose is to eliminate the manual effort users spend monitoring trading activities by automating whale discovery, analysis, and trade replication. The system actively discovers new whale wallets, evaluates their trading performance, and maintains a curated list of high-performing traders. When these tracked whales execute trades, an AI agent analyzes multiple factors including portfolio composition, market data, and relevant news to determine if the trade should be replicated. For users who opt in, successful trades are automatically executed across their connected wallets, with optimizations for each user's specific balance and conditions.

Solution

Data Collection Layer: Uses TheGraph with a custom subgraph to identify large transactions and discover whales. Alchemy webhooks monitor whale wallet activities in real-time. AI Processing Layer: Built on CDP AgentKit and LangChain, utilizing OpenAI models. Custom tools integrate Brave Search, Alchemy SDK (for wallet analysis), and cryptocurrency APIs (CoinGecko/CoinCodex) for market data. Backend Infrastructure: NestJS server hosted on Autonome, handling agent operations and Alchemy Address Activity webhooks processing use Smart Contract Layer: Deployed on Base Mainnet, implements a trading wallet factory pattern. Frontend: NextJS application with OnChainKit for Web3 interactions, featuring fuzzy search and market cap weighted token search to enhance LLM analysis accuracy. Execution Layer: Uses ethers and viem for blockchain interactions, with Basescan API integration for contract verification and transaction analysis.The system operates through three main processes: whale discovery (daily CRON jobs analyzing on-chain data), trade analysis (real-time evaluation of whale transactions), and trade execution (automated replication across user wallets with position sizing optimization).