Getting Started

Set up a Python 3.12-3.14 environment and install the package:

pip install -U --pre nautilus_trader

PyPI publishes NautilusTrader 2.x as 2.0.0rcN pre-releases. Include --pre until 2.0.0 is released. Confirm the install with:

python -c "import nautilus_trader; print(nautilus_trader.__version__)"

Installation

Platform support, package indexes, and building from source.

Quickstart

Run your first backtest in five minutes using synthetic data.

Choose your path

Backtesting: learn the two API levels below, then work through the tutorials for strategy pattern walkthroughs.
Live trading: see the Configure a live trading node how-to and Integrations for supported venues.
Data workflows: see the how-to guides for loading external data and setting up the Parquet data catalog.
Building adapters: see the Developer guide.

Backtesting API levels

NautilusTrader provides two API levels for backtesting:

API level	Entry point	Best for
Low-level	`BacktestEngine`	Direct component access, library development
High-level	`BacktestNode`	Production workflows, easier transition to live trading (recommended)

The high-level API requires a Parquet-based data catalog. The low-level API works with in-memory data but has no live-trading path.

One node per process

Running multiple BacktestNode or LiveNode instances concurrently in the same process is not supported due to global singleton state. Sequential execution with proper disposal between runs is supported. A replacement LiveNode on the same thread also requires dropping the previous node, or releasing all references to it in Python, before construction. See Processes and threads.

See the Backtesting guide for help choosing an API level.

Backtest (low-level API)

Backtest an EMA cross strategy with a TWAP execution algorithm on Binance trade ticks with BacktestEngine.

Backtest (high-level API)

Load raw data into the data catalog, then run a backtest with BacktestNode.

Examples

The docs cover a subset of examples. For the full set, see the nautilus_trader repository.

Directory	Description
examples/	Runnable Python examples organized by environment
docs/tutorials/	Python and Rust tutorials for common workflows
docs/concepts/	Concept guides with code snippets
python/tests/unit/	Unit tests covering core functionality and edge cases

Running in Docker

A self-contained Jupyter notebook server is available as a Docker image, with no local setup required.

docker pull ghcr.io/nautechsystems/jupyterlab:latest --platform linux/amd64
docker run -p 8888:8888 ghcr.io/nautechsystems/jupyterlab:latest

Then open http://localhost:8888 in your browser.

Container data is ephemeral; deleting the container removes all data.

NautilusTrader log output can exceed Jupyter's default rate limit, causing notebooks to hang. Set LoggerConfig(stdout_level=LogLevel.ERROR) to avoid this.