NOFX is an AI trading terminal that lets traders run language-model-driven strategies across US stocks, commodities, forex, and cryptocurrency markets.
The tool addresses the challenge of deploying AI agents in live trading by wrapping language model decisions in a Go runtime that enforces hard risk limits no model can override. Traders define a strategy once, then the system enters a continuous loop: read market data, let the model decide, execute the trade, and record the reasoning. The approach separates strategy logic from risk control, so a model cannot accidentally or intentionally breach position limits, leverage caps, or other guardrails the trader sets.
NOFX suits traders who want to experiment with AI-driven strategies without building infrastructure from scratch. The tool supports nine exchanges including Hyperliquid, Binance, and traditional stock brokers, so you can test the same strategy across different markets. Everything runs locally on your machine; exchange credentials are encrypted at rest and never transmitted. The project includes a public leaderboard where traders can compare realized returns across different strategies and models running in parallel. Getting started requires only registering an account, funding an AI fee wallet with USDC on Base, and connecting an exchange.
Development activity shows consistent engagement with the codebase, regular updates addressing user-reported issues, and active maintenance of core trading and risk-management logic. The project maintains documentation in multiple languages and sustains a community channel for user discussion and feedback.