Shimmy

What it is and what it does

Shimmy is a bridge library that adapts external reinforcement learning environments to work with Gymnasium and PettingZoo, two popular standardized RL APIs. It solves the problem of environment fragmentation: RL researchers and practitioners often need to work with simulators built for different APIs, but want to use them within a modern, consistent framework. Shimmy wraps these external environments, translating their interfaces to match Gymnasium or PettingZoo conventions.

The package depends on numpy and gymnasium as its core runtime dependencies. It requires Python 3.10 or later and is actively maintained, with support for Python versions through 3.14. The library is designed for researchers, practitioners, and teams building RL agents who need to integrate multiple environment sources without rewriting integration code for each one.

Use it for:

• Adapt a legacy Gym-based environment to work with modern Gymnasium code without rewriting your agent
• Use multiple third-party RL simulators within a single unified Gymnasium interface
• Integrate external multi-agent environments with PettingZoo for centralized agent training
• Build RL benchmarks that test agents across diverse environment sources without environment-specific glue code
• Migrate existing RL codebases from older APIs to Gymnasium by wrapping environments instead of refactoring

Worth the install?

AI-flagged interpretation of the facts on this page — verify before relying

Shimmy converts external reinforcement learning environments to Gymnasium and PettingZoo APIs, letting you use popular RL simulators with a unified interface.

Yes. Shimmy is actively maintained, has no known vulnerabilities, uses a permissive MIT license, and solves a real fragmentation problem in RL development. Low install friction and support for current Python versions make it a practical choice for anyone integrating multiple RL environments.

Install

shimmy on PyPI

pip install shimmy

uv add shimmy

poetry add shimmy

Installing Shimmy

Before you install

Low friction installation as a pure-Python wheel. Actively maintained with a recent commit on 2026-08-08 and regular releases; the project is not archived and has accumulated 222 repository stars.

License in practice

MIT license permits commercial and private use with minimal restrictions—you may use, modify, and distribute Shimmy freely provided you retain the license notice.

Quickstart

pip install shimmy

import shimmy
import gymnasium

# Shimmy provides bindings to wrap external RL environments
# for use with Gymnasium and PettingZoo APIs

Requires Python 3.10 or later; numpy and gymnasium must be installed as runtime dependencies.

Verify before relying

• Which specific external RL environments are currently supported by version 2.0.1
• Whether Shimmy handles environment state serialization or checkpointing
• Performance overhead of the API conversion layer compared to native environment calls

Package facts

Evidence: shimmy-2.0.1-py3-none-any.whl

Keywords: Reinforcement Learning, game, RL, AI

Tags

More Artificial Intelligence packages