tensortrade-org/tensortrade

What it is

An open-source Python framework for building, training, and evaluating reinforcement learning agents for algorithmic trading. It provides composable components—environments, action schemes, reward schemes, observers, and data feeds—that can be combined to create custom trading systems, with training scripts built around Ray RLlib and Optuna. Traders and researchers experimenting with RL-based strategies will find its tutorials, architecture guides, and documented experiments useful. The project's own research notes that commission costs can exceed prediction profits for frequently trading agents, so results depend heavily on assumptions like trading frequency and fees.