FinRL-Meta is the market environment and benchmark layer of the FinRL ecosystem, building a universe of financial market environments for data-driven reinforcement learning. It enables users to develop, evaluate, and compare deep reinforcement learning trading agents across diverse financial tasks. The project provides dynamic datasets and uses OpenAI Gym-style environments to simulate market conditions for agent training.