Skip to content

Technical Clarification: MediaPipe Telemetry Configuration for iOS and Android #6397

Open

Description

@chandrakeerthim

Have I written custom code (as opposed to using a stock example script provided in MediaPipe)

None

OS Platform and Distribution

iOS 18 & Android 12

MediaPipe Tasks SDK version

iOS: 1.0.0; Android: 0.10.32

Task name (e.g. Image classification, Gesture recognition etc.)

Face Landmarker

Programming Language and version (e.g. C++, Python, Java)

Swift (iOS), Kotlin (Android)

Describe the actual behavior

MediaPipe Tasks SDK automatically collects and transmits telemetry data when used for Face Landmarker processing. As discussed in GitHub issue #6291, telemetry is enabled by default, and there is no official API to disable it. We are looking for clarification on the telemetry behaviour and available options for both iOS and Android SDKs.

Describe the expected behaviour

We would like to understand how telemetry collection can be disabled or prevented on both iOS and Android while continuing to use Face Landmarker functionality. We would like Google's recommendation on the supported approaches, including blocking the telemetry endpoint or building the SDK from source without the telemetry component.

Standalone code/steps you may have used to try to get what you need

Not applicable. This is a request for technical clarification regarding MediaPipe SDK telemetry behaviour and configuration across iOS and Android platforms. Reference: #6291

Other info / Complete Logs

We reviewed the discussion in GitHub issue #6291, where a Google engineer confirmed that telemetry is enabled by default and there is no official opt-out API.

We also reviewed the latest MediaPipe v1.2.2 release notes and noticed that v1.2.0 introduced Web Worker platform detection and Flutter Web telemetry overrides.

Could you please clarify whether any equivalent telemetry configuration or opt-out API is available for native iOS and Android applications in v1.2.2?

If no native API is available, are blocking the telemetry endpoint or building the SDK from source still the recommended approaches?

We would appreciate guidance on the feasibility and potential impact of these approaches on FaceLandmarker functionality across both platforms.

Reference: https://github.com/google-ai-edge/mediapipe/issues/6291

Latest release: https://github.com/google-ai-edge/mediapipe/releases/tag/v1.2.2
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions