What it does

Qbot is a locally-deployed quantitative investment platform that tries to cover the full lifecycle: data ingestion, strategy backtesting, paper trading, and live execution for stocks, funds, futures, and crypto. It bundles a GUI (desktop with some web views), a backtest engine, and hooks for machine-learning strategies. Notifications can ping you via email, Feishu, WeChat, or pop-ups when your bot decides to trade.

The interesting bit

The project treats “factor mining” as a first-class citizen—automated discovery of predictive signals, backtested before admission to a factor library. It also explicitly layers in supervised learning, reinforcement learning, and deep learning as strategy paradigms, not just technical indicators. That ambition is rare in retail-grade open-source quant tools.

Key highlights

• Modular architecture: data layer, strategy layer, and trade engine are abstracted so you can swap data sources or brokers.
• Supports multiple asset classes: stocks, funds, futures, and virtual currencies.
• Includes a model zoo and multi-factor workflow for AI-driven strategies.
• Simulated trading with latency and slippage approximation before going live.
• Python 3.8/3.9 only; other versions untested.

Caveats

• The README is a mix of Chinese and English, and some sections (like the strategy pool description) are truncated mid-sentence.
• “Only tested on Python 3.8/3.9” is a narrow window for a project with this many dependencies.
• The live trading integrations (vnpy, pytrader) are mentioned but not documented in detail in the provided source; expect glue-code archaeology.

Verdict

Worth a look if you are a Python-literate retail trader curious about ML-driven quant strategies and willing to tinker with a partially-documented stack. If you need battle-tested, institution-grade infrastructure or clear English documentation, this is not your tool yet.