Crazyflow

Overview

Crazyflow is a high-performance research simulator for Crazyflie‑style small quadrotors.
Built on JAX and MuJoCo, it supports batched GPU execution, differentiable dynamics, and accurate, identified models — designed for reproducible experiments at scale.

Audience: researchers working on control, learning, system identification, sim2real, multi-agent RL and swarm control for quadrotors.

Highlights

Modular simulation stack (physics, integrator, controller)
GPU-ready, batched execution for massive parallelism
Differentiable / autodiff-enabled dynamics
Support for analytical (identified) and data-driven models
Onboard-controller support and symbolic model matching
Extensible step pipeline (randomization, disturbances, custom hooks)
MuJoCo-based visualization and offscreen rendering

Performance benchmarks and comparisons

Quick start

Installation — follow the install instructions: Installation Guide
Run an example — see runnable demos and thumbnails: Examples
API & reference — full Python API generated with mkdocstrings: API Reference

Want to learn more?

Read about features and architecture: Features & Architecture
Try the simple quickstart: Getting Started / Usage