Shadow

Shadow is an AI-driven portfolio management and trading assistant designed for the typical web3 user. It functions as a “shadow” by continuously monitoring on-chain activity and market sentiment—even while the user is busy or asleep—and provides actionable insights. In Phase 1, the agent passively recommends tokens based on analysis of on-chain trading habits, social sentiment, and technical data. In Phase 2, the agent actively executes trades based on these insights while incorporating user and “smart money” wallet data.

Features: Automated Portfolio Management: Provide users with real-time insights and recommendations for token trades based on historical and current market data. User Convenience: Allow users to “set it and forget it” by having the agent monitor trends and even execute trades on their behalf. Smart Money Analysis: Leverage data from notable “smart money” wallets to provide additional perspective and recommendations. Market Sentiment and Technical Analysis Integration: Combine social sentiment (from X.com) and technical indicators (from Dexscreener) to provide a holistic trading strategy.

Built using LangChain's framework with the following APIs integrated:

1. Farcaster for sentiment
2. Dexscreener for TA
3. Solscan + basescan for wallet transactions

Agent model using GPT 4o currently but is easily swappable.

Terminal and message queue built using RabbitMQ.

No partner technologies used.