# simpler-env/SimplerEnv

Evaluating and reproducing real-world robot manipulation policies (e.g., RT-1, RT-1-X, Octo) in simulation under common setups (e.g., Google Robot, WidowX+Bridge) (CoRL 2024)

Repository: https://github.com/simpler-env/SimplerEnv
Canonical: https://ross.abutalabs.com/products/simplerenv
Homepage: https://simpler-env.github.io/
Language: Jupyter Notebook
License: MIT
License Family: permissive
Topics: computer-vision, embodied-ai, real2sim, reinforcement-learning, robot-learning, robot-manipulation, robotics, robotics-simulation, robotics-benchmark
Last push: 2025-12-20T13:08:29+00:00

## Health v2 (maintenance only)
Score: 51/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 58, release rhythm 35, longevity 63
- inputs: {"age_days": 893, "days_push": 256, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1147, forks 198 (observed 2026-08-28T04:03:45.838025+00:00)

## What it is
SIMPLER (SimplerEnv) is a collection of simulated environments built on SAPIEN/ManiSkill for evaluating real-world robot manipulation policies like RT-1, RT-1-X, and Octo. It provides real-to-sim evaluation setups (Visual Matching and Variant Aggregation) that correlate strongly with real-world robot performance.

## Use cases
- evaluate robot manipulation policies in simulation
- reproduce real-world robot policy results without physical robots
- benchmark RT-1 or Octo policies on Google Robot and WidowX setups
- select checkpoints for robot learning models using sim metrics
- study real-to-sim gap for manipulation evaluation
- run scalable reproducible robot policy evaluations

## When to choose
- you need to evaluate generalist manipulation policies without expensive real-robot runs
- you want standardized, reproducible robot evaluation setups
- you need sim-based checkpoint selection or failure mode analysis for robot policies

## When to avoid
- you need full-fidelity digital twins of specific real environments
- your robot setup or task is not among the supported environments
- you need high-fidelity contact dynamics rather than policy-level evaluation

## Facets
- artifact type: framework
- maturity: active
- function: simulation, machine-learning, computer-vision, benchmarking, robotics
- domain: robotics, machine-learning, simulation, computer-vision, artificial-intelligence
- platform: python
- tags: robot-manipulation, real2sim, robot-learning, maniskill, sapien, policy-evaluation, benchmark, embodied-ai, reinforcement-learning, linux, gpu

## Member repositories
- simpler-env/SimplerEnv (main) score 51

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:45.838025+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:34:03.921672+00:00, confidence not recorded.
  - readme: https://github.com/simpler-env/SimplerEnv (fetched 2026-08-28T04:03:45.838025+00:00, sha 49ca5a69a5a9)
  - homepage: https://simpler-env.github.io/ (fetched 2026-08-29T12:39:21.727652+00:00, sha 277e38bde704)
- Data as of 2026-08-30T08:39:29.467469+00:00.