Skip to content
RobothecaBeta
Log in

SimplerEnv

  • Software
  • Humanoids and robot learning
  • Tool

A simulation benchmark for evaluating robot manipulation policies against real-world setups, built on SAPIEN and ManiSkill2 with Bridge dataset environments in ManiSkill3.

Made by simpler-env.

Source on GitHub: github.com/simpler-env/SimplerEnv (opens in a new tab). Facts read at commit eddb569 (opens in a new tab).

Licenses

Code
MIT (opens in a new tab)

Read by Robotheca from LICENSE (opens in a new tab) (MIT) at commit eddb569 (opens in a new tab), on 16 Oct 2026.

Commercial use
Yes

What it does

  • Provides simulated environments that mirror real Google Robot and WidowX tasks, such as picking a coke can or stacking cubes.
  • Offers a visual matching setup that overlays real-world images onto simulation backgrounds to align appearances.
  • Offers a variant aggregation setup that averages results across sim variants with different backgrounds, lighting, distractors and textures.
  • Includes example inference scripts for RT-1, RT-1-X and Octo policies, and a minimal reset-and-step loop to start from.
  • Provides real and simulated performance data for policies on all tasks, plus functions for computing MMRV and Pearson correlation metrics.
  • A GPU-parallelized version of the Bridge environments in ManiSkill3 is available on a separate branch and runs 10-15x faster.

Written by Robotheca from the project's README (opens in a new tab) at commit eddb569.

Getting started

Difficulty
Intermediate
Hardware it needs, as the project states it
An NVIDIA GPU (ideally RTX; non-RTX GPUs such as 1080Ti and A100 are slower on ray-tracing environments), CUDA version >=11.8 and <13 for full installation, and a CPU-only ManiSkill2 path; TPUs are not supported because SAPIEN requires a GPU.

Maintenance

Status
Active: a commit in the last 90 days
Newest commit
30 Sept 2026 on main
Latest release
None published on GitHub

Checked on GitHub on 16 Oct 2026.

Links

Robots on Robotheca that use it

No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.

Robotheca lists software as a reference: nothing is hosted or run here, and every fact links to where it was read.