LIQWID - Autonomous Trading Agent

LIQWID is an autonomous trading agent designed to turn structured market analysis into disciplined, explainable trading decisions. The system combines a strategy engine, conviction scoring, risk management, automated execution, and an MCP-based tool interface into a single paper-trading pipeline. LIQWID continuously retrieves market data, scans for predefined trading opportunities, evaluates signal strength, applies risk controls such as position sizing and allocation limits, and can execute approved trades through Alpaca's Paper Trading API. Every decision and action is logged, creating an auditable history of what the system observed, decided, and executed. The MCP architecture allows the same trading capabilities to be accessed through a CLI or connected to an LLM agent, creating a foundation for an AI system that can reason over market state and interact with controlled trading tools rather than directly accessing the broker. The project is currently designed for paper trading and strategy validation. Our initial backtesting has produced encouraging early results, but we are continuing to test the strategy across a larger sample and different market conditions before making claims about statistical profitability. LIQWID's goal is to demonstrate how AI, automation, structured decision-making, and safety-oriented risk controls can work together to create a transparent and controllable autonomous trading system.