MLKitRuntime XCFramework

Reproducible dynamic XCFramework wrapper around the ML Kit modules used by OneLightApps Flutter applications. OCR and image-labeling models are embedded inside their owning framework so host applications need no resource-copy build phases.

Build

./scripts/build_xcframework.sh

Output: build/output/MLKitRuntime.xcframework.zip.

Runtime dependencies

The host Swift package must link these products once:

• GTMSessionFetcherCore 3.5.0
• GULLogger and GULUserDefaults 8.1.0
• GoogleDataTransport 10.1.0
• GoogleToolboxForMac from the matching ML Kit 9.0.0-1 binaries

These stay outside MLKitRuntime so applications that already use Firebase do not receive a second embedded copy of the shared Google Objective-C classes.

Architectures

• iOS device: arm64
• iOS Simulator: x86_64

The upstream ML Kit 9.0.0-1 binaries do not contain an arm64 simulator slice.