FinRL: Financial Reinforcement Learning Framework for Automated Trading

FinRL: Financial Reinforcement Learning Framework for Automated Trading

Summary

FinRL is the first open-source framework for financial reinforcement learning, providing an ecosystem for automated trading in quantitative finance. It offers a comprehensive pipeline, various DRL algorithms, and support for multiple market environments and data sources, making it a powerful tool for researchers and practitioners.

Repository Info

Updated on October 12, 2025
View on GitHub

Tags

financial reinforcement learningalgorithmic tradingdeep reinforcement learningfintechstock tradingDRL frameworkJupyter NotebookAI in finance

Click on any tag to explore related repositories

Introduction

FinRL is a pioneering open-source framework dedicated to financial reinforcement learning, offering a robust ecosystem for automated trading in quantitative finance. Developed by the AI4Finance-Foundation, FinRL provides a comprehensive, three-layered architecture encompassing market environments, intelligent agents, and diverse financial applications. With over 12,800 stars and 2,900 forks on GitHub, it serves as a vital resource for researchers and practitioners exploring the intersection of AI and finance.

Installation

Getting started with FinRL is designed to be straightforward. You can install the library via pip:

pip install finrl

For detailed installation instructions across various operating systems, including macOS, Ubuntu, and Windows 10, please refer to the official documentation: FinRL Installation Guide.

Examples

FinRL provides numerous tutorials and examples to help users understand and implement deep reinforcement learning strategies for financial tasks. A quick start can be found in Stock_NeurIPS2018.ipynb, demonstrating a basic stock trading application. Key tutorials include:

Why Use FinRL?

FinRL stands out as a leading platform for several reasons:

  • First Open-Source Framework: It was the first open-source framework for financial reinforcement learning, establishing a benchmark in the field.
  • Comprehensive Ecosystem: FinRL has evolved into a broader ecosystem, including projects like FinRL-Meta for gym-style market environments and ElegantRL for advanced DRL algorithms.
  • Automated Pipeline: It offers an automatic train-test-trade pipeline, simplifying the development and deployment of trading strategies.
  • Extensive Data Support: The framework supports a wide array of data sources, including YahooFinance, Alpaca, Binance, and many more, providing flexibility for diverse financial markets.
  • Robust Algorithms: Integrates popular DRL libraries such as Stable Baselines3 and Ray RLlib, allowing users to leverage state-of-the-art algorithms.
  • Strong Academic Backing: Supported by numerous publications in top-tier conferences and journals, demonstrating its scientific rigor and practical applicability.

Links