Documentation

Getting Started

QuantCraft is an all-in-one desktop app for developing, testing, and iterating Python trading strategies.
This guide covers only what you need to get running quickly and correctly. For the big picture, read the Introduction; for IDE regions and panels, see Workspace layout.

Prerequisites

Install these before using strategy features:

  • QuantCraft desktop app (latest stable version)
  • Python 3.10+ (3.11 recommended)
  • pip (included with most Python installs)
  • Internet access (for package installs and connected data/services)
  • Basic disk space for projects, logs, and backtest outputs
  • Virtual environments (venv) for dependency isolation
  • A project-level requirements.txt for reproducible installs

1. Install Python

Download and install Python from python.org.

During install:

  • Enable Add Python to PATH

Verify:

python --version pip --version

If python is not recognized on Windows, try:

py --version

2. Install and Open QuantCraft

  • Install QuantCraft using your standard installer/package
  • Launch QuantCraft
  • Open the strategy IDE/workspace area

3. Create Your First Strategy

  1. Create or import a .py strategy file
  2. Add core lifecycle callbacks:
    • on_init
    • on_bar (and/or on_tick, on_timer)
    • on_finish
  3. Define configurable params for tunable values

4. Run a Backtest

Configure:

Use Test (running code), then review:

See Backtests for how the engine drives your strategy.


5. Confirm Setup Is Healthy

Your setup is correct when:

  • Python scripts execute without import/runtime errors
  • Backtests complete and produce result panels
  • Strategy logs and charts match expected behavior

Common Setup Issues

  • python not found: reinstall Python and enable PATH
  • Module import errors: install dependencies in the active environment
  • Inconsistent results: verify symbol, date range, timeframe, and params
  • No trades generated: check entry/exit conditions and warmup/data availability

Best Practices for New Users

  • Start with one symbol and a simple strategy
  • Parameterize constants (don’t hardcode everything)
  • Log key decisions for debugging
  • Validate on multiple date ranges/market regimes before trusting results