Home / Backtesting & trading frameworks
FinRL
Deep reinforcement learning library for automated quantitative trading
What it is
An open-source framework for financial reinforcement learning, built around a train-test-trade pipeline for automated trading research. Agents such as A2C, DDPG, PPO, SAC and TD3 can be trained and backtested using Gym-style market environments and data processors wired to sources like Yahoo Finance. Developers, learners and researchers prototyping DRL strategies in Jupyter Notebook form the core audience. The codebase is aimed at education, benchmarking and research prototyping rather than production deployment, and live trading support is limited to basic Alpaca integration.
At a glance
Worth watchingOur rating, based on popularity, maintenance and how ready it is for real use.
| Best for | Developers |
|---|---|
| Used for | Backtesting, Strategy research |
| Markets | Multi-market |
| Stack | Python |
| Learning curve | Moderate learning curve |
| Practical value | Medium practical value |
| Cost | Free and open source |
| Hardware | GPU optional |
| Maintenance | Last commit 2 days ago |
GitHub stars, last 30 days
Daily snapshots since 2026-09-12 (up to 30 days): +342 over the period, now 16,620. Gaps mean no snapshot was taken that day.
A Deep Reinforcement Learning Library for Automated Trading in Quantitative Finance. NeurIPS 2020.
Similar tools
Python algorithmic trading framework with machine learning support
Python backtesting and live trading platform for strategies
Fast limit order book simulator for trading strategy research
Python reinforcement learning framework for training trading agents
Modular schedule-driven backtesting engine for equities and ETFs
Gym-style environment for reinforcement learning trading and backtesting