Habitat
A modular high-level library to train embodied AI agents across a variety of tasks and environments. observed · 2026-08-28
Health v2 · maintenance only
74/100
- Activity 81
- Release rhythm 51
- Longevity 100
How is this computed?
round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-03. Adoption (stars, forks) is never an input.
- gap_med: 276.5
- age_days: 2767
- days_rel: 118
- days_push: 118
- n_releases_24m: 3
Adoption not part of the score
3108 stars · 687 forks observed · 2026-08-28
What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded
AI Habitat is a high-performance, physics-enabled 3D simulation platform for Embodied AI research, consisting of Habitat-Sim (a fast 3D simulator with sensors, robots, and rigid-body dynamics) and Habitat-Lab (a modular library for training agents on embodied tasks). It renders photorealistic indoor/outdoor scenes at thousands of FPS and supports datasets like HM3D, MatterPort3D, Gibson, and Replica.
Use cases
- train embodied AI agents for navigation in photorealistic 3D indoor scenes
- simulate a Fetch robot rearranging objects with rigid-body physics
- benchmark reinforcement learning agents on embodied navigation and instruction-following tasks
- render RGB-D observations from 3D scans like MatterPort3D or HM3D at high frame rates
- test sim-to-real transfer of robot policies before deploying to hardware
- simulate mobile manipulators and quadrupeds described via URDF
When to choose
- you need extremely fast rendering and physics stepping for large-scale RL training of embodied agents
- you want built-in support for standard embodied AI datasets (HM3D, MatterPort3D, Gibson, Replica, ReplicaCAD)
- you are doing Embodied AI research such as navigation, rearrangement, or human-robot interaction
- you need configurable sensors and URDF-described robots in a Python-friendly simulator
When to avoid
- you need a simulator under active official development and support - Meta has stopped maintaining it beyond v0.3.4
- you need broad general-purpose simulation features over raw speed (the design prioritizes FPS over capability breadth)
- you need outdoor vehicle dynamics or soft-body physics simulation
- you want a game-engine-quality rendering pipeline with modern PBR features
Facets
library · maturity maintenance
simulation machine-learning computer-vision robotics graphics artificial-intelligence robotics computer-vision simulation machine-learning python cpp embodied-ai 3d-simulator sim2real reinforcement-learning-environments physics-simulation meta-research linux macos gpu
8 sources
- readme: https://github.com/facebookresearch/habitat-lab · fetched 2026-08-28 · 94b028249038
- homepage: https://aihabitat.org/ · fetched 2026-08-29 · 5e47ebbf258e
- site_page: https://aihabitat.org/docs/resource_index · fetched 2026-08-29 · f7a8472d15a4
- site_page: https://aihabitat.org/docs/concepts · fetched 2026-08-29 · 07d72c57c791
- site_page: https://aihabitat.org/docs/habitat-sim · fetched 2026-08-29 · 0d0e2a9b936b
- site_page: https://aihabitat.org/docs/habitat-lab · fetched 2026-08-29 · 0718a4868ddd
- site_page: https://aihabitat.org/docs · fetched 2026-08-29 · 6a27d7f82bd2
- registry_pypi: https://pypi.org/pypi/habitat-lab/json · fetched 2026-08-29 · 363257a31ae8
Member repositories
| Repository | Role | Health v2 |
|---|---|---|
| facebookresearch/habitat-lab | main | 74 |
| facebookresearch/habitat-sim | backend | 75 |
For agents
markdown · JSON · MCP: product_card(name="facebookresearch/habitat-lab")
Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem