Genesis World

Health v2 · maintenance only

Adoption not part of the score

30058 stars · 2887 forks observed · 2026-10-15

What it is AI-extracted, prompt v1, taxonomy v1, 2026-10-05, confidence 0.97

Genesis World is a Python‑based simulation platform that combines a unified multi‑physics engine, a photorealistic renderer, and a cross‑platform compiler. It enables researchers to build, run, and scale embodied AI and robotics experiments from a laptop to datacenter GPUs.

Use cases

• train robot control policies in a high‑fidelity simulated environment
• generate synthetic sensor data (camera, lidar, depth) for computer‑vision research
• benchmark and compare different physics models for manipulation tasks
• run large‑scale parallel simulations for reinforcement‑learning pipelines
• prototype and test new robotic hardware designs before physical prototyping
• create photorealistic datasets for domain‑randomization and sim‑to‑real transfer

When to choose

• You need a single Pythonic API that integrates physics, rendering, and compilation.
• Your workloads must scale from CPU‑only laptops to multi‑GPU clusters.
• You require support for multiple asset formats (URDF, MJCF, OBJ, GLB, USD).
• You want an open‑source, actively maintained platform with a permissive Apache‑2.0 license.

When to avoid

• You require a real‑time engine with sub‑millisecond latency for hardware‑in‑the‑loop testing.
• Your project depends on a specific game‑engine ecosystem (e.g., Unity or Unreal).
• You need a compiled binary distribution without a Python runtime dependency.

Facets

library · maturity active

simulation compiler image-processing artificial-intelligence simulation python windows cross-platform robotics embodied-ai multi-physics photorealistic-renderer gpu-scaling linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Genesis-Embodied-AI/genesis-world")