# Habitat

A modular high-level library to train embodied AI agents across a variety of tasks and environments.

Repository: https://github.com/facebookresearch/habitat-lab
Canonical: https://ross.abutalabs.com/products/habitat
Homepage: https://aihabitat.org/
Language: Python
License: MIT
License Family: permissive
Topics: ai, computer-vision, robotics, simulator, sim2real, deep-learning, deep-reinforcement-learning, reinforcement-learning, research, python
Last push: 2026-05-07T22:03:51+00:00
Link (homepage): https://aihabitat.org/
Link (site_page): https://aihabitat.org/docs/resource_index
Link (site_page): https://aihabitat.org/docs/concepts
Link (site_page): https://aihabitat.org/docs/habitat-sim
Link (site_page): https://aihabitat.org/docs/habitat-lab

## Health v2 (maintenance only)
Score: 74/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 81, release rhythm 51, longevity 100
- inputs: {"age_days": 2767, "days_push": 118, "days_rel": 118, "gap_med": 276.5, "n_releases_24m": 3}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3108, forks 687 (observed 2026-08-28T04:07:43.821134+00:00)

## What it is
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
- artifact type: library
- maturity: maintenance
- function: simulation, machine-learning, computer-vision, robotics, graphics
- domain: artificial-intelligence, robotics, computer-vision, simulation, machine-learning
- platform: python, cpp
- tags: embodied-ai, 3d-simulator, sim2real, reinforcement-learning-environments, physics-simulation, meta-research, linux, macos, gpu

## Member repositories
- facebookresearch/habitat-lab (main) score 74
- facebookresearch/habitat-sim (backend) score 75

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:43.821134+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 2026-08-29T18:26:54.921942+00:00, confidence not recorded.
  - readme: https://github.com/facebookresearch/habitat-lab (fetched 2026-08-28T04:07:43.821134+00:00, sha 94b028249038)
  - homepage: https://aihabitat.org/ (fetched 2026-08-29T09:21:58.848069+00:00, sha 5e47ebbf258e)
  - site_page: https://aihabitat.org/docs/resource_index (fetched 2026-08-29T09:21:58.857521+00:00, sha f7a8472d15a4)
  - site_page: https://aihabitat.org/docs/concepts (fetched 2026-08-29T09:21:58.859620+00:00, sha 07d72c57c791)
  - site_page: https://aihabitat.org/docs/habitat-sim (fetched 2026-08-29T09:21:58.861534+00:00, sha 0d0e2a9b936b)
  - site_page: https://aihabitat.org/docs/habitat-lab (fetched 2026-08-29T09:21:58.863358+00:00, sha 0718a4868ddd)
  - site_page: https://aihabitat.org/docs (fetched 2026-08-29T09:21:58.864985+00:00, sha 6a27d7f82bd2)
  - registry_pypi: https://pypi.org/pypi/habitat-lab/json (fetched 2026-08-29T09:21:58.866513+00:00, sha 363257a31ae8)
- Data as of 2026-08-30T08:39:29.467469+00:00.