nautechsystems/nautilus_trader

NautilusTrader is an open-source trading engine designed for building and running automated trading strategies at a professional level. It covers the full workflow: researching and testing a strategy using historical data, then deploying that same strategy live with real money, all without rewriting the code. The core problem it addresses is a common pain point in algorithmic trading. Most traders develop strategies in Python using simplified, data-centric tools, then must rewrite everything in a faster compiled language when going live. NautilusTrader eliminates this gap by providing a Rust engine underneath that handles the speed and reliability requirements, while Python sits on top as the language you use to write your actual trading logic. The same strategy code works in both research simulation and live production. The system uses an event-driven architecture, meaning it reacts to market events (price updates, order fills, etc.) as they happen, rather than running in batch loops. This is how professional trading systems work because it handles the real-world timing and ordering of events correctly. Historical backtesting uses the same architecture, so simulated results closely reflect what live execution would actually do. It supports many asset types, crypto exchanges, foreign exchange, equities, futures, options, and includes adapters for connecting to various trading venues. The engine is written in Rust with Python bindings, runs on Linux, macOS, and Windows, and is fast enough to also train AI trading agents.