Overview
NautilusTrader is an open-source trading engine from Nautech Systems of Sydney. A Rust core runs the event loop, and strategies are written in Python or Rust. The pitch is one codebase for research, backtesting and live trading, with no rewrite when a strategy goes live.
It is a library you run yourself, not a data feed. It loads market data into a Parquet catalog and connects to venues and vendors such as Databento, Tardis and Interactive Brokers through adapters. A hosted Pro service and Cloud platform are announced but not yet priced.
Good to know
- It supplies no market data of its own. You pay your data vendor separately and load the data into the catalog.
- The 2.0 line is still release candidates: the latest, 2.0.0rc9, is dated 10 Nov 2026. PyPI's latest stable is 1.231.0 (2 Aug 2026), and installing 2.x needs pip's --pre flag.
- Python version support differs between pages (docs say 3.13 or 3.14 for install; the repo README lists 3.13-3.15; PyPI 1.231.0 lists 3.12-3.14), so check before installing.
- Pro, Cloud and the dashboard are not generally available and show no prices. The LGPL-3.0 license and a trademark policy limit use of the name and logo by third-party projects.
Latest news
Screens


Pricing
Core engine is free and open source. Paid Pro tiers are limited by number of live nodes; no prices published as of 21 Nov 2026. Free use: Yes (Open-source engine is free under LGPL-3.0. Pro (waitlist) free plan monitors and controls one NautilusTrader node. Data vendors charge separately.). Prices as listed on 21 Nov 2026.
What you get
Backtest-live parity
The same strategy code runs in simulation and in production on shared engine semantics.
Rust-native core
Core components are written in Rust, with a Python strategy API and the option of pure Rust systems.
Nanosecond simulation
Event-driven backtests at nanosecond resolution with configurable fill, fee, latency and order book models.
Order book depth
The simulated exchange accepts L1, L2 and L3 book data and supports ten order types.
Parquet data catalog
Stores market data as Parquet (Arrow record batches) for replay, research and backtesting.
Rich data types
Built-in types include order book deltas and depth, quotes, trades, bars, mark and index prices, funding rates and option greeks.
Bar aggregation
18 aggregation methods including tick, volume, value, imbalance, runs, Renko and time bars.
Advanced orders
Post-only, reduce-only, iceberg, OCO, OUO and OTO orders with IOC, FOK, GTC, GTD and DAY time in force.
Multi-venue adapters
Modular adapters connect brokers, exchanges and data vendors to the same engine.
Data coverage
21 integrations listed (18 stable, 3 beta): venues and brokers plus data providers Databento, Tardis, CoinAPI, FinFeed. It brings no data of its own, so history depth depends on the vendor you connect.
Works with
History
Trader reviews
How we score →No reviews on Fintech Rankings yet. Be the first: we read every review before it appears.
Write the first reviewSimilar tools
All data and devtools →Questions
How much does NautilusTrader cost?
The engine is free under LGPL-3.0. Pro and Cloud have no published prices and are on a waitlist or in development as of 21 Nov 2026.
Can I use it for free?
Yes, the open-source engine is free to use. Pro's free plan lets you monitor and control one NautilusTrader node. You still pay any data vendor or broker you connect.
Which markets and venues does it cover?
Its docs list 21 integrations, including Binance, Bybit, OKX, Coinbase, Kraken, Hyperliquid, dYdX, Interactive Brokers, Betfair, Polymarket and Kalshi, plus data vendors Databento, Tardis and CoinAPI.
How much history does it have?
It holds none itself. History depth comes from the data vendor you load into its Parquet catalog; no depth figure is stated on its pages.
Is it historical, live or both?
Both. The same strategy code runs in nanosecond-resolution backtests and live with connected venues.
Can I use it commercially or redistribute it?
The code is LGPL-3.0-or-later, which allows use under that license's terms. The name and logo are trademarked; third-party projects must follow the trademark policy and logo use needs written permission.

