Essentials

Immersive features

Updated: Apr 6, 2026

Meta Quest provides a set of spatial computing features that let your app understand and interact with the physical world. These capabilities transform VR from a screen-replacement technology into a spatially aware platform that responds to the user’s environment, body, and gaze.

Scene understanding

Scene understanding gives your app a 3D model of the user’s physical space. The system detects walls, floors, ceilings, furniture, and other surfaces, providing your app with a semantic map of the room.

Key capabilities:

Room layout — Walls, floor, and ceiling boundaries

Furniture detection — Tables, couches, desks, and other objects identified with bounding boxes

Semantic labels — Each detected element is labeled (wall, floor, table, couch, window, door)

Mesh generation — Detailed 3D meshes of the physical environment for collision and occlusion

Plane detection — Flat surfaces identified for placing virtual content

Scene understanding enables apps that adapt to the user’s room — a game that builds a castle out of your couch, a fitness app that marks safe exercise zones, or a design tool that lets you preview furniture in your actual living room.

Spatial anchors

Spatial anchors let you pin virtual content to specific locations in the physical world. When a user places a virtual painting on their wall, a spatial anchor ensures it stays on that wall across sessions — even after the user restarts their headset.

Key capabilities:

Persistent anchors — Virtual content remains in place across app sessions

Shared anchors — Multiple users see the same virtual content at the same physical location

Cloud anchors — Anchors stored in the cloud for cross-device persistence

Anchor tracking — Real-time position and orientation updates as the headset moves

Spatial anchors are essential for any mixed reality app where content should feel physically attached to the real world. Without anchors, virtual objects drift and lose their spatial relationship to the environment.

Body tracking

Body tracking estimates the user’s full-body pose from headset and controller data. This enables expressive avatars, body-driven gameplay, and fitness applications that track movement accuracy. Full body tracking is available through both engine-specific SDKs and Meta Spatial SDK.

Key capabilities:

Full-body estimation — Head, torso, arms, and legs estimated from headset and controller positions

Avatar driving — Map body pose to virtual avatars in real time

Gesture detection — Recognize full-body gestures like jumping, ducking, or reaching

Fitness tracking — Track body movement for exercise and wellness applications

Runtime fidelity control — Adjust tracking quality (HIGH or MEDIUM) based on your app’s needs

Joint velocities — Access velocity data for each tracked joint

Body tracking accuracy is highest for the upper body (head, arms) where sensor data is direct, and estimated for the lower body (legs, feet) through inverse kinematics. Design your app to work well with upper-body data and treat lower-body estimation as supplemental.

Note: The OS may dynamically reduce body tracking fidelity under thermal pressure (for example, dropping from HIGH to MEDIUM). Design your app to handle graceful quality degradation.

Eye tracking

Eye tracking is available on Meta Quest Pro and detects where the user is looking. This enables gaze-based interactions, social eye contact in multiplayer experiences, and performance optimizations through foveated rendering. Quest 3 and Quest 3S do not have eye tracking hardware.

Key capabilities (Quest Pro only):

Gaze direction — 3D vector indicating where each eye is looking

Eye openness — Detect blinks and eye closure

Eye-tracked foveated rendering — Render at full resolution only in the user’s gaze direction, significantly reducing GPU workload

Social presence — Drive avatar eye movement for natural eye contact in multiplayer

Eye tracking requires user consent. Always provide fallback behavior for users on devices without eye tracking (Quest 3, Quest 3S).