# NVlabs/ProtoMotions

ProtoMotions is a GPU-accelerated simulation and learning framework for training physically simulated digital humans and humanoid robots.

Repository: https://github.com/NVlabs/ProtoMotions
Canonical: https://ross.abutalabs.com/products/protomotions
Homepage: https://nvlabs.github.io/ProtoMotions/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: character-animation, digital-human, humanoid, humanoid-robots, physics-simulation, reinforcement-learning
Last push: 2026-08-20T01:12:02+00:00

## Health v2 (maintenance only)
Score: 66/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 98, release rhythm 35, longevity 50
- inputs: {"age_days": 708, "days_push": 14, "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 2338, forks 465 (observed 2026-08-28T04:06:38.042029+00:00)

## What it is
ProtoMotions is a GPU-accelerated simulation and reinforcement learning framework for training physically simulated digital humans and humanoid robots. It provides a modular, multi-backend prototyping platform (Newton, IsaacGym, IsaacLab, Genesis, MuJoCo) for physics-based character animation and humanoid robotics research.

## Use cases
- train physically simulated humanoid characters with reinforcement learning
- learn motion skills from the AMASS motion capture dataset
- simulate and control humanoid robots like Unitree G1 and H1
- prototype physics-based character animation research
- train motion imitation policies like MaskedMimic, AMP, and ASE
- retarget motion capture data to robot morphologies
- generate procedural terrains for humanoid training

## When to choose
- you need GPU-accelerated large-scale humanoid RL training
- you want to swap between multiple physics simulators with one codebase
- you work across animation and humanoid robotics communities
- you need state-of-the-art motion imitation algorithms out of the box

## When to avoid
- you need a lightweight CPU-only setup for simple physics tasks
- you want a production animation tool rather than a research framework
- your project does not involve humanoid characters or RL

## Facets
- artifact type: framework
- maturity: active
- function: machine-learning, reinforcement-learning, simulation
- domain: machine-learning, robotics, simulation, graphics, deep-learning
- platform: python, cross-platform
- tags: humanoid, character-animation, digital-humans, motion-imitation, isaac-gym, isaaclab, newton, mujoco, genesis, motion-retargeting, robotics-research, gpu, linux

## Member repositories
- NVlabs/ProtoMotions (main) score 66

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:38.042029+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 2026-08-30T02:37:30.383790+00:00, confidence not recorded.
  - readme: https://github.com/NVlabs/ProtoMotions (fetched 2026-08-28T04:06:38.042029+00:00, sha b76073c68b40)
  - homepage: https://nvlabs.github.io/ProtoMotions/ (fetched 2026-08-29T10:18:07.090469+00:00, sha 893f5004e7ad)
- Data as of 2026-08-30T08:39:29.467469+00:00.