Telemetry

Updated: Jun 23, 2026

Overview

This page describes the telemetry data that the Device Access Toolkit SDK collects from Android and iOS apps that integrate with Meta AI glasses.
The SDK collects operational and diagnostic telemetry (for example, device discovery, registration flows, streaming session lifecycle, permission checks, attestation, and performance markers) to support reliability, performance monitoring, troubleshooting, and product improvement. In turn, telemetry helps Meta identify issues, measure SDK health, and improve the developer experience.
Some of the data collected includes:
  • Device identifiers
  • Firmware versions
  • Session durations
  • Error types
  • Success/failure flags
Supported devices for telemetry: Ray-Ban Meta glasses, Oakley Meta HSTN, Oakley Meta Vanguard, and Meta Ray-Ban Display (MRBD).

Opt Out

Telemetry collection is enabled by default, but you can disable it at the app level.

Android

Add the following to your AndroidManifest.xml inside <application>:
<meta-data
    android:name="com.meta.wearable.mwdat.ANALYTICS_OPT_OUT"
    android:value="true" />

iOS

Add the following to your Info.plist:
<key>MWDAT</key>
<dict>
    <key>Analytics</key>
    <dict>
        <key>OptOut</key>
        <true/>
    </dict>
</dict>

Data Categories

We collect the following categories of telemetry data via the SDK:

Application and SDK identification

Identifies which third-party app and SDK version generated an event, enabling usage attribution and version-specific issue diagnosis:
Examples:
  • App bundle ID / package name, app name, and app version
  • Public-key SHA-256
  • App session UUID and session start time, which resets on each app start
  • SDK version (for example, 0.7.0)

Device and platform context

Describes the phone and glasses involved in an interaction.
Examples:
  • Phone OS, OS version, and device model
  • Glasses device identifier, device type (for example, RAYBAN_META)
  • Glasses SoC firmware build version
  • Platform and link mode (dynamic, static, or mixed)

Registration and authentication

Tracks the one-time app linking flow between your app and the Meta AI app.
Examples:
  • Registration step (for example, started, completed, failed)
  • App linking flow ID (UUID for end-to-end correlation)
  • Attestation session ID and event type

Permissions

Records when your app checks or requests device capabilities through the Meta AI app.
Examples:
  • Permission name (for example, CAMERA, MICROPHONE) and whether it is granted
  • Success or failure of the permission check/request API, with error details if applicable
  • For bulk queries, a map of permission names to consent states

Session lifecycle

Measures the connection lifecycle between your app and the glasses. A session represents a sustained period of access to device sensors or outputs.
Examples:
  • Session state transitions: starting, started, paused, stopping, stopped
  • Previous session state
  • Error details if a transition fails

Camera streaming

Captures usage and quality signals for video and photo streaming.
Examples:
  • Stream event type (prepare started/completed, start, stop, error)
  • Configuration: video codec, audio codec, resolution tier (LOW 360×640, MEDIUM 504×896, HIGH 720×1280), frame delivery method
  • Duration in seconds

Display content delivery

For glasses with display capabilities (Meta Ray-Ban Display glasses), logs display session usage.
Examples:
  • Display event type (start, stop, send content, video playback states)
  • Display session ID and duration
  • Video source type, video codec, payload size in bytes
  • Video transfer duration in milliseconds

Device discovery

Tracks glasses availability from the SDK’s perspective.
Examples:
  • Event type (device_discovered, device_forgotten, error)
  • Device identifier
  • Error details if applicable

Error reporting and diagnostics

Supports stability monitoring and crash triage.
Examples:
  • Structured error name, description, and type category
  • Crash grouping hash, crash type (signal or exception class)
  • Breadcrumbs (last N analytics event names before the crash)

Developer testing with Mock Device Kit

When you use Mock Device Kit to test without physical hardware, separate events are logged.
Examples:
  • Mock device event type, mock device ID, device type code
  • Mock service event type, service ID, success/error details