TradePilot is a desktop app that trains machine-learning models on ticker patterns, scans a watchlist, makes automatic paper trades, and tracks fake profit/loss before any real money is involved.
Trains a model for each ticker, ranks probability, and paper-trades automatically.
Trains a model for each ticker and estimates upside probability from recent market features.
Runs the bot while the app is open, buying and selling paper positions based on model output. Real broker trades are planned for a later safety-locked mode.
You set the maximum account percentage allowed. The bot decides how much of that max to actually use.
Shows probability, confidence, model accuracy, RSI, trend, volume ratio, and top features.
Tracks fake cash, holdings, open P/L, realized P/L, and trade history.
First launch includes a local account and an app-style tutorial that highlights key controls.
TradePilot can automatically paper-trade while the app is open. It trains ML models, scans tickers, chooses entries/exits, sizes positions, and logs every fake trade reason.
A future Paper / Real Money switch can connect to a broker API, but it should stay locked until the paper bot proves it can handle losses, drawdowns, and risk limits.
TradePilot is paper-trading only right now. The goal is to test the bot through backtesting, live paper trading, drawdown tracking, and repeated tuning before even considering real trades.
The goal is not magic or guaranteed wins. The goal is controlled losses, stronger average wins, better rules, and enough testing to know when the model is not ready.
Download buttons are placeholders while the app is still being packaged. The current build runs as a native desktop app locally.