Open-Source Framework for Trading Crypto with Large Models 1. NOFX Project (NoFx) NOFX is a general architecture AI trading operating system built on large language models like DeepSeek and Qwen. It’s an open-source implementation and replica of the viral AI quantitative trading system NOF1. This framework aims to provide a complete closed-loop system, covering everything from decision-making to trading and review. Large Model Integration and Features: NOFX has achieved AI task flow support, forming a closed loop of Context → Large Model → Decision Results → Trade Execution. Supported Exchanges: It currently supports three major exchanges—Binance, Hyperliquid, and Aster DEX—with one-click switching. Deployment and Community: Supports one-click deployment via Docker or manual installation. Some users and developers are actively using and optimizing the nofx AI trading logic, such as reducing trading frequency to minimize frequent stop-losses. Core Features: The project includes key features like partial close + dynamic take-profit/stop-loss, multi-strategy template switching (e.g., default/nof1/adaptive), and Telegram news integration (for real-time market sentiment analysis). Official Account (X): @nofx_ai Project Website: https://(nofxai.com) GitHub Repository: http://(((github.com)))/tinkle-community/nofx GitHub PR (Updates): https://(((github.com)))/NoFxAiOS/nofx/pull/220 Trading Logic Guide: https://(((github.com)))/ShenXuGongZi/easyNOFX/blob/main/TRADING_LOGIC_GUIDE.md If traditional quantitative trading systems are like race cars that require precise instructions, then open-source frameworks like NOFX, based on large models, are like equipping the car with an “AI driver” (Large Model Agent) that can think, learn, and self-optimize. It can automatically make trading decisions and manage risks based on higher-dimensional information (e.g., pre-set strategies or market sentiment).