FinRL

FinRL is built for learning and researching financial reinforcement learning. It packages environments, agents, training, and backtesting into one workflow, with a basic path for connecting trained strategies to Alpaca.

reinforcement learningstrategy backtestingautomated tradingeducation and research
GitHub Stars15,859
Primary languageJupyter Notebook
LicenseMIT
Official repositoryOpen GitHub ↗

Is it a fit?

The quick decision view, grounded in the ranking data and verification record.

Best for

Researchers who want to reproduce or modify reinforcement-learning trading experiments in Python or notebooks.

Not ideal for

People who only need a lightweight rules-based backtest or do not plan to train models.

Tested status

Not locally tested

Setup difficulty

Medium

Last verified

2026-07-31

What it can do

Research and testing

Strategy backtesting
Yes
LLM or agent based
No

Trading

Paper trading
Not confirmed
Automated trading
Yes

Getting started

API key required to start
No
Registration required to start
No
Self-hostable
Yes

Capabilities shown as Yes or No come from the official project source. Unconfirmed fields remain “Not confirmed”.

Related task rankings

The same tool can rank differently for different jobs.

Best AI Research ToolsAI Trading IndexRank 5 →Best Quant Strategy Discovery ToolsAI Trading IndexRank 6 →Best Strategy Backtesting ToolsAI Trading IndexRank 10 →Best Paper Trading ToolsAI Trading IndexRank 8 →Best Automated Trading ToolsAI Trading IndexRank 6 →Best Overall AI Trading ToolsAI Trading IndexRank 7 →