RunAnywhere v0.0.1-dev.ee5013d

RunAnywhere v0.0.1-dev.ee5013d

On-device ML inference framework for iOS/macOS/Android

iOS Installation

Swift Package Manager:

dependencies: [
    .package(url: "https://github.com/RunanywhereAI/runanywhere-binaries.git", from: "0.0.1-dev.ee5013d")
]
targets: [
    .target(name: "YourApp", dependencies: ["RunAnywhereCore"])
]

CocoaPods:

pod 'RunAnywhereCore', '~> 0.0.1-dev.ee5013d'

The unified XCFramework includes all backends (ONNX Runtime + LlamaCPP + WhisperCPP) and supports both iOS and macOS.

macOS Installation

macOS apps use the same XCFramework as iOS. However, ONNX Runtime is a dynamic library on macOS and must be embedded in your app bundle:

1. Download onnxruntime-macos.zip from this release
2. Extract and copy libonnxruntime*.dylib to your app's Frameworks/ directory
3. In Xcode: Target → General → Frameworks → Add → Embed & Sign

Alternatively, install via Homebrew: brew install onnxruntime

Android Installation

Gradle (with RunAnywhere Kotlin SDK):

// Add to your module's build.gradle.kts
dependencies {
    implementation("com.runanywhere:runanywhere-core-onnx:0.0.1-dev.ee5013d")
}

The SDK module will automatically download pre-built native libraries.

Direct Download:
Download RunAnywhereONNX-android.zip and extract to your project's jniLibs directory.

Capabilities

• Text Generation - Run LLMs locally on device
• Embeddings - Generate text embeddings for semantic search
• Speech-to-Text - Batch and real-time streaming transcription
• Text-to-Speech - Natural voice synthesis
• Voice Activity Detection - Detect speech in audio
• Speaker Diarization - Identify different speakers

Platform Support

Platform	Architecture	Notes
iOS Device	arm64	Static library
iOS Simulator	arm64 + x86_64	Static library
macOS	arm64 + x86_64	Static library + ONNX dylib
Android	arm64-v8a	Shared library (.so)

Verification

shasum -a 256 -c checksums.txt

---

Built from runanywhere-core @ ee5013d8fc8d73196988c8bed8502162ceebd35a