Nautilus, also known as NautilusTrader, is a free, open-source engine for building and running algorithmic trading systems. It is aimed at developers, quantitative researchers and technically minded traders who want to write their own strategies, test them on historical data and run the same logic in live markets. The project is licensed under LGPL, and its latest release is v2.0.0rc8, a release candidate.
The engine is designed for multi-asset, multi-venue trading, so one system can work with several instruments across more than one exchange. Its core is written in Rust, which the project chose for speed and safety. Expect to write strategies and configuration yourself, because Nautilus is a framework rather than a finished bot. The most recent code update is dated 2026-10-31.
What Is Nautilus and Who Is It For?
Nautilus, short for NautilusTrader, is a software engine for building and running algorithmic trading systems. The project describes itself as production-grade, and its core is written in Rust, a language chosen for speed and memory safety. It handles multi-asset, multi-venue setups, meaning one system can trade different kinds of instruments across more than one exchange or trading venue.
The people it suits are mostly developers, quantitative researchers and technically minded traders. If you want to write a strategy in code, test it against historical data and then run the same logic on a live account, the engine covers that whole loop. If you are looking for a ready-made bot with a point-and-click interface, the alternatives listed below may fit better.
Teams exploring machine learning trading agents may also find the engine useful, because its simulation layer is fast enough to generate large numbers of training episodes.
How Event-Driven Backtesting Works
An event-driven backtester does not simply apply a formula to a price chart. It replays the market in the order events happened: a quote arrives, a trade prints, a bar closes, and the strategy reacts to each one in turn. This sequence matters because it mirrors what a live system would have seen at each moment, which makes the results more realistic than a shortcut calculation over closing prices.
Nautilus feeds its backtester with several data types, including quote ticks, trade ticks, bars, order book data and custom series. Resolution is measured in nanoseconds, so small timing differences between events can be represented faithfully. A single simulation can include several venues and instruments at once, which is useful when a strategy depends on price differences between exchanges.
A backtest is only as good as its data. Gaps in history or badly aligned timestamps can make a strategy look better or worse than it would have been. Check your data sources carefully before trusting any result, and remember that a good backtest does not guarantee future performance.
Key Concepts: Venues, Adapters, Cache and Message Bus
Several terms come up often in the documentation. A venue is any place where instruments trade, such as a cryptocurrency exchange. An adapter translates a venue's REST API or WebSocket feed into the engine's internal format, so strategies do not need to know the details of each exchange's protocol.
The cache is the engine's shared store of recent market data, orders and positions. Components read from it instead of querying exchanges repeatedly, which keeps the system responsive. The message bus is the channel through which components publish and subscribe to events, so a strategy, a risk check and a logger can each react to the same update without being tightly coupled.
Understanding these pieces makes the documentation much easier to follow. It also helps when deciding whether a piece of logic belongs inside a strategy or in a separate component that other parts of the system can reuse.
Key features
Rust-Powered Speed
Type and Thread Safety
Runs on Windows, macOS and Linux
Modular Adapters
User-Defined Components
Event-Driven Backtesting
Same Code for Research and Live Trading
Multi-Venue Strategies
AI Training Support
Supported coins
Nautilus is built for multi-asset, multi-venue trading, which means it is designed to handle different instruments across several exchanges or venues within one setup. The exact list of supported markets depends on the adapters you use, so check the documentation for the specific venue you plan to trade on before building a strategy around it.
How to install Nautilus on Windows
- 1
Get the Release from the Official Source
Go to the project's GitHub repository at github.com/nautechsystems/nautilus_trader and open the Releases page. Download only from the official website or the official GitHub Releases page. The latest version is v2.0.0rc8, a release candidate, so read its release notes before you start.
- 2
Verify the File
Check the SHA-256 checksum of the downloaded file against the value published with the release. If a signature is provided, verify it as well. A mismatch means the file should not be installed, and you should download it again from the official source.
- 3
Install or Unpack the Windows Build
Run the Windows installer or unpack the Windows build, depending on what the release provides. Install into a folder you control and avoid running the program from a shared or temporary location. Keep your system updated and make sure you have enough free disk space for data files.
- 4
Optional: Deploy with Docker
If you prefer a containerized setup, install Docker on Windows and follow the Docker deployment instructions in the documentation. Docker keeps the engine and its dependencies isolated from the rest of your system, which makes it easier to reproduce the same environment later.
-
Check the Installation
Open the documentation at nautilustrader.io/docs/latest and run the smallest example it provides. Confirm that the example completes and produces the expected output before connecting any exchange account or building a larger system.
How to use Nautilus
Working with Nautilus is a developer workflow rather than a click-through setup. Start with the documentation and a small example, then add complexity step by step.
- Read the documentation at nautilustrader.io/docs/latest and run the smallest example it includes.
- Choose the instruments and venues you want to work with, and use the adapter for each exchange.
- Write your strategy as a component that reacts to incoming data events and decides when to place or cancel orders.
- Load historical data for the period you want to test, and check that timestamps, formats and symbols are correct.
- Configure the backtest with your venues, instruments and strategies, then run it.
- Review the results, including fees, slippage assumptions and the number of trades, before changing any parameters.
- Repeat the test on a different time period to see whether the strategy still behaves sensibly, rather than tuning it to one stretch of data.
- Only after thorough testing, consider a limited live deployment with small position sizes and close monitoring.
Pros & cons
- Free and open-source under the LGPL license
- Rust core designed for speed and type safety
- Same strategy code for backtesting and live trading
- Event-driven backtests with nanosecond resolution
- Supports multi-venue and cross-venue strategies
- Runs on Windows, macOS and Linux, and can be deployed with Docker
- Modular adapters for REST APIs and WebSocket feeds
- Engine is described as fast enough for AI agent training
- Requires programming skills and a willingness to write strategies yourself
- The framework has a learning curve, and the documentation is essential reading
- Latest version v2.0.0rc8 is a release candidate rather than a final release
- Standard-precision mode uses 64-bit integers with up to 9 decimals and a smaller value range
- Not a plug-and-play bot for users who want a ready-made interface
Is Nautilus safe?
Nautilus is open-source, which lets anyone inspect the code, but safety still depends on how you get the software and how you use it. Download only from the official website or the official GitHub Releases page, and verify the SHA-256 checksum or signature before running any file. Never enter a seed phrase, private key or exchange password on a third-party website. When you connect exchange API keys, grant only the permissions your strategy needs and disable withdrawals where the exchange allows it. The last code update is dated 2026-10-31, and the master branch reflects the latest released version and is recommended for production use. The nightly branch contains daily snapshots of the develop branch for early testing, so treat it as experimental.
- Official source only. Download from github.com/nautechsystems/nautilus_trader. Ignore ads and lookalike domains.
- Check the SHA-256. Compare the installer hash with the value the developer publishes before running it.
- Never share your seed phrase. No developer, support agent or website — including CryptApps — will ever ask for it.
Alternatives
Nautilus shares some ground with other tools in this catalog, but its focus is different. Jesse is also a framework for strategies and backtests, so it is a natural comparison for developers. The distinguishing points of Nautilus are its Rust core and its nanosecond-resolution, multi-venue simulation. Elinesoft XTester and TestMyLogic put more emphasis on checking strategies through simulation or testing them for overfitting. Nautilus aims to cover research, backtesting and live trading within one engine. For users who prefer a web interface, OctoBot is closer to an out-of-the-box experience. Gunbot and Moon Bot are local trading bots, and Moon Bot is popular in the CIS region for working with Binance and other exchanges.
Specifications
- Price
- Free
- Type
- Open source
- License
- LGPL
- Latest version
- v2.0.0rc8
- Last update
- 31 Oct 2026
- Windows support
- Yes (Windows build or installer)
- Source code
- github.com/nautechsystems/nautilus_trader
Frequently asked questions
Yes, Nautilus is free and open-source, licensed under LGPL. You can download the Windows build or installer without paying for the software itself. Costs you may still face include exchange fees on trades, data feeds, server hosting if you run the engine remotely, and your own time spent learning the framework and building strategies.
Yes, Nautilus runs on Windows, and the project supports a Windows build or installer. The engine also runs on Linux and macOS, and it can be deployed using Docker. Before you install, check the release notes for the version you plan to use, since the current release is a release candidate and details may change in later versions.
In practice, yes. Nautilus is a framework for building and running trading systems, so you will write strategies, configure components and set up backtests yourself. Basic programming skills and a clear understanding of order types, market data and exchange APIs will make the documentation much easier to follow and will reduce mistakes during testing.
Yes, backtesting is one of its main purposes. The event-driven backtester replays historical quote ticks, trade ticks, bars, order book data and custom data at nanosecond resolution. You can run multiple venues, instruments and strategies in one simulation, which helps you test ideas before deciding whether they deserve more attention or real capital.
The project describes the engine as fast enough to train AI trading agents using reinforcement learning or evolution strategies. Training usually requires a very large number of simulated steps, so simulation speed is a practical advantage. Results from training still depend on the quality of the data, the reward design and the way the agent is evaluated.
The master branch reflects the source code of the latest released version and is the one the project recommends for production use. The nightly branch contains daily snapshots of the develop branch, merged at 14:00 UTC and as needed. It is meant for early testing, so it may contain changes that have not yet been fully validated.
Several similar programs are available in this catalog. Jesse is a framework for strategies and backtests, while Elinesoft XTester focuses on simulation and strategy testing. TestMyLogic checks strategies for overfitting, and OctoBot offers a web interface. The right choice depends on whether you need a research framework, a ready-made bot or a dedicated testing tool.
Verdict
Nautilus is a strong choice for developers and quantitative researchers who want one engine for backtesting, simulation and live trading, with a Rust core built for speed. It suits people who are comfortable writing code and reading technical documentation. It is a poor fit for anyone who wants a plug-and-play bot. Since the latest version is a release candidate, test thoroughly and consider the master branch for production work. Check the official website for the current release before you install.