Jesse Framework: Beginner Crypto Bot Guide

freqtrade is the most popular open-source crypto bot, but it is not the only one. Jesse is a newer framework that targets traders who want more advanced backtesting, cleaner code architecture, and machine-learning integration. It is particularly popular among crypto traders who have outgrown simpler tools.

This guide introduces Jesse's core concepts and walks through a first strategy setup.

What Is Jesse?

Jesse is a Python framework for designing, backtesting, and deploying algorithmic trading strategies on cryptocurrency markets. It emphasizes:

• Performance: Backtest engine optimized for speed and accuracy.
• Machine learning: Built-in support for ML-driven strategies.
• Modularity: Clean separation between routes, strategies, and research tools.
• Live trading: Direct exchange integration for automated execution.

Jesse is often compared to freqtrade, but it targets a slightly more advanced user base.

Quick Comparison: Jesse vs freqtrade

If you want a simple bot with lots of community strategies, freqtrade is probably better. If you want a clean, modern framework with strong ML support, Jesse is worth exploring.

Installation

Jesse is installed via pip. It is recommended to use a dedicated Python environment.

python -m venv jesse-env
source jesse-env/bin/activate
pip install jesse

After installation, create a new project:

jesse make-project my-trading-project
cd my-trading-project

This scaffolds the directory structure with config, strategies, storage, and research folders.

Project Structure

A typical Jesse project looks like this:

my-trading-project/
├── config.py
├── routes.py
├── strategies/
│   └── MyStrategy.py
├── storage/
└── research/

• config.py: Exchange API keys, symbols, timeframes, and backtest settings.
• routes.py: Maps symbols to strategies.
• strategies/: Contains your strategy classes.
• storage/: Holds imported historical data.
• research/: Notebooks and experiments.

A Simple Strategy

Here is a basic trend-following strategy in Jesse:

from jesse.strategies import Strategy
 
class TrendFollowing(Strategy):
    def should_long(self):
        return self.close > self.sma(50)
 
    def should_short(self):
        return False
 
    def go_long(self):
        qty = self.capital / self.close
        self.buy = qty, self.close
 
    def update_position(self):
        if self.close < self.sma(50):
            self.liquidate()

Jesse strategies read like a set of rules. The framework handles execution, logging, and reporting.

Backtesting

To run a backtest, first import historical data:

jesse import-candles Binance BTC-USDT 1d

Then run the backtest:

jesse backtest

Jesse provides detailed reports including equity curve, drawdown, trades, and metrics such as Sharpe ratio and win rate.

Machine Learning Integration

Jesse supports ML strategies through its research module. A typical workflow is:

1. Extract features from historical data.
2. Train a model using scikit-learn, XGBoost, or a neural network.
3. Save the model and load it inside the strategy.
4. Use model predictions as signals.

This integration is more native than in many other frameworks, making Jesse attractive for ML-curious traders.

Live Trading

Jesse supports live trading through exchange adapters. Before going live:

• Test extensively in backtest and paper modes.
• Use API keys with restricted permissions.
• Set up monitoring and alerts.
• Start with small position sizes.

When to Choose Jesse

Choose Jesse if:

• You want a modern Python framework with clean architecture
• Machine learning is central to your strategy
• You value backtesting performance
• You are comfortable with a smaller but growing community

When to Choose freqtrade Instead

Choose freqtrade if:

• You want the largest community and most public strategies
• You prefer a built-in web UI
• You want broader exchange support through CCXT
• You are newer to algorithmic trading

Common Pitfalls

• Over-optimizing ML models: Financial data is noisy. A model that backtests well often fails live.
• Ignoring fees: Crypto exchange fees vary widely and can erase edges.
• Over-leverage: Futures strategies can produce large losses fast.
• Data gaps: Incomplete historical data leads to unrealistic backtests.

Bottom Line

Jesse is a powerful alternative to freqtrade for traders who want advanced backtesting and machine-learning support in a clean Python framework. It requires more setup and learning than beginner tools, but the payoff is a more flexible and performant platform.

If you are already comfortable with Python and want to level up your crypto algo trading, Jesse deserves a serious look.

Related reading: Freqtrade Beginner Guide | Freqtrade vs Hummingbot | Top AI Trading GitHub Projects