A large language model trading agent running autonomously on Hyperliquid. It reads on-chain data, interprets market microstructure, and executes trades — all for $1.30/day in API costs.
Reads funding rates, open interest, liquidation zones, and order book depth from Hyperliquid's on-chain data in real-time.
The model interprets market state as a language problem — analyzing patterns the same way it processes natural language.
Combines quantitative signals (entropy, Hurst, FVG) with LLM interpretation to generate high-conviction trading signals.
Executes trades on Hyperliquid with position sizing, risk management, and stop-loss logic — no human intervention required.
Markets are not physics problems to be solved with differential equations. They're language problems — complex, context-dependent, and best understood by models that can process nuance and ambiguity. LLMs are the first models sophisticated enough to read markets the way humans do, but at machine speed.
Read the full thesis →
Why treating financial markets as natural language problems — not physics problems — yields better predictive models. The theoretical framework behind our AI trading approach, inspired by Renaissance Technologies and modern NLP.
How to use on-chain perpetual futures data — funding rates, open interest, and liquidation heatmaps — as quantitative trading signals. Python implementations and backtest results on crypto markets.
How we built an autonomous LLM-powered trading agent that runs on Hyperliquid for $1.30/day in API costs. Complete architecture, cost breakdown, prompt engineering, and live results.