#trading #backtesting #investment #crypto #stock


Framework for building event-driven live-trading & backtesting engines

51 releases

new 0.8.15 Jul 12, 2024
0.8.13 Apr 16, 2023
0.8.12 Feb 18, 2023
0.8.6 Dec 3, 2022
0.1.0 Mar 27, 2021

#1244 in Magic Beans

Download history 18/week @ 2024-03-15 18/week @ 2024-03-22 44/week @ 2024-03-29 31/week @ 2024-04-05 2/week @ 2024-04-12 8/week @ 2024-04-19 5/week @ 2024-04-26 1/week @ 2024-05-24 9/week @ 2024-05-31 4/week @ 2024-06-07 2/week @ 2024-06-14 7/week @ 2024-06-21 172/week @ 2024-06-28

189 downloads per month

MIT license

6.5K SLoC


Barter is an open-source Rust framework for building event-driven live-trading & backtesting systems. Algorithmic trade with the peace of mind that comes from knowing your strategies have been backtested with a near-identical trading Engine. It is:

  • Fast: Barter provides a multi-threaded trading Engine framework built in high-performance Rust (in-rust-we-trust).
  • Easy: Barter provides a modularised data architecture that focuses on simplicity.
  • Customisable: A set of traits define how every Barter component communicates, providing a highly extensible framework for trading.

See: Barter-Data, Barter-Integration & Barter-Execution

Crates.io MIT licensed Build Status Discord chat

API Documentation | Chat


Barter is an open-source Rust framework for building event-driven live-trading & backtesting systems. It provides a high-performance, easy to customise trading Engine that enables backtesting strategies on a near-identical system to live trading. The Engine can be controlled by issuing Commands over the Engine's command_tx. Similarly, the Engine's Events can be listened to using the event_rx (useful for event-sourcing). At a high level, it provides several de-coupled components that interact via a set of traits:

  • Data: MarketGenerator trait governs the generation of a MarketEvents that acts as the system heartbeat. For example, a live::MarketFeed implementation is provided that utilises Barter-Data WebSocket integrations to provide live exchange data (ie/ trades, candles, etc).
  • Strategy: The SignalGenerator trait governs potential generation of Signal after analysing incoming MarketEvents. Signals are advisory and sent to the Portfolio for analysis.
  • Portfolio: MarketUpdater, OrderGenerator, and FillUpdater govern global state Portfolio implementations. A Portfolio may generate OrderEvents after receiving advisory SignalEvents from a Strategy. The Portfolio's state updates after receiving MarketEvents and FillEvents.
  • Execution: The ExecutionClient trait governs the generation of FillEvents after receiving OrderEvents from the Portfolio. For example, a SimulatedExecution handler implementation is provided for simulating any exchange execution behaviour required in dry-trading or backtesting runs.
  • Statistic: Provides metrics such as Sharpe Ratio, Calmar Ratio, and Max Drawdown to analyse trading session performance. One-pass dispersion algorithms analyse each closed Position and efficiently calculates a trading summary.
  • Trader: Capable of trading a single market pair using a customisable selection of it's own Data, Strategy & Execution instances, as well as shared access to a global Portfolio.
  • Engine: Multi-threaded trading Engine capable of trading with an arbitrary number of Trader market pairs. Each contained Trader instance operates on its own thread.


  • For brevity: Imports are not included - see /examples for everything you need!
  • For simplicity:
    • Engine and Trader(s) are configuration with hard-coded values rather than loaded in configuration values.
    • Engine only contains one Trader (usually you would have many Traders, one for each Market).
    • Remote Commands (eg/ Command::Terminate, Command::ExitPosition) are not sent to the Engine via it's command_tx, this control over the Engine can be added as per your taste (eg/ connected to an HTTP endpoint).
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Setup Logger & Load Config For Engine & Trader Instances Here

    // Create channel to distribute Commands to the Engine & it's Traders (eg/ Command::Terminate)
    let (command_tx, command_rx) = mpsc::channel(20);
    // Create Event channel to listen to all Engine Events in real-time
    let (event_tx, event_rx) = mpsc::unbounded_channel();
    let event_tx = EventTx::new(event_tx);
    // Generate unique identifier to associate an Engine's components
    let engine_id = Uuid::new_v4();
    // Create the Market(s) to be traded on (1-to-1 relationship with a Trader)
    let market = Market::new("binance", ("btc", "usdt", InstrumentKind::Spot));
    // Build global shared-state MetaPortfolio (1-to-1 relationship with an Engine)
    let portfolio = Arc::new(Mutex::new(
            .allocation_manager(DefaultAllocator { default_order_value: 100.0 })
            .risk_manager(DefaultRisk {})
            .statistic_config(StatisticConfig {
                starting_equity: 10_000.0,
                trading_days_per_year: 365,
                risk_free_return: 0.0,
            .expect("failed to build & initialise MetaPortfolio"),
    // Build Trader(s)
    let mut traders = Vec::new();
    // Create channel for each Trader so the Engine can distribute Commands to it
    let (trader_command_tx, trader_command_rx) = mpsc::channel(10);

            .strategy(RSIStrategy::new(StrategyConfig { rsi_period: 14 }))
            .execution(SimulatedExecution::new(ExecutionConfig {
                simulated_fees_pct: Fees {
                        exchange: 0.1,
                        slippage: 0.05,
                        network: 0.0,}
            .expect("failed to build trader")
    // Build Engine (1-to-many relationship with Traders)
    // Create HashMap<Market, trader_command_tx> so Engine can route Commands to Traders 
    let trader_command_txs = HashMap::from_iter([(market, trader_command_tx)]);
    let engine = Engine::builder()
        .statistics_summary(TradingSummary::init(StatisticConfig {
            starting_equity: 1000.0,
            trading_days_per_year: 365,
            risk_free_return: 0.0
        .expect("failed to build engine");
    // Listen to Engine Events & do something with them
    // --- Run Trading Session Until Remote Shutdown OR Data Feed ends naturally (ie/ backtest) ---

Getting Help

Firstly, see if the answer to your question can be found in the API Documentation. If the answer is not there, I'd be happy to help to Chat and try answer your question via Discord.


🎉 Thanks for your help in improving the barter ecosystem! Please do get in touch on the discord to discuss development, new features, and the future roadmap.

In addition to the Barter crate, the Barter project also maintains:

  • Barter-Integration: High-performance, low-level framework for composing flexible web integrations.
  • Barter-Data: High performance & normalised WebSocket integration for leading cryptocurrency exchanges - batteries included.
  • Barter-Execution: In development and soon to be integrated.


  • Integrate the new "Barter-Execution" library to provide additional out-of-the-box functionality for executing OrderEvents.
  • Build Strategy utilities for aggregating tick-by-tick Trades and/or Candles into multi-timeframe datastructures, as well as providing an example multi-timeframe Strategy using the utilities.
  • Flesh out the MetaPortfolio Allocator & Risk management systems.
  • And more!


This project is licensed under the MIT license.


Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in Barter by you, shall be licensed as MIT, without any additional terms or conditions.


~420K SLoC