ML Lab is still in development, so join now for early access.
From hypothesis to a model in your strategy.
ML Lab models can feed the visual builder through an ml_predict node. Walk-forward replay uses real prediction series when available and warns when they are missing.
Access market data
Start from structured market data across asset classes.
Train your model
Train a machine-learning, statistical or stochastic model, whichever suits the idea.
Backtest and validate
Replay the model's available predictions walk-forward against history.
Deploy to a strategy
Add the trained model to a strategy as an ml_predict node.
See it in action
Pick a ticker, timeframe, and lookback, then stack preset indicators or custom transforms and tune their weights with a live preview.
Statistical, stochastic and machine-learning models.
Train ML models
Train models from gradient-boosted trees to neural networks without setting up any infrastructure.
Statistical models
Fit ARIMA, GARCH and Markov-style models to market data.
Stochastic models
Simulate Monte Carlo price paths for probabilistic forecasts.
Market data access
Work with data for selected equity, forex and crypto markets.
Backtest trained models
Replay predictions walk-forward when a trained series is available.
Plug into strategies
Use a model as a signal, a filter or an execution layer in a strategy.
Feature engineering
Build model inputs from pre-built indicators or custom transforms.
Guardrails and validation
Inspect training results and the warnings a backtest raises.