Xybrid
SDKs

Android

Native Android SDK for on-device ML inference

The Android SDK provides native Kotlin/Java bindings to the Xybrid runtime for Android applications.

Coming Soon - The Android SDK is currently in development. In the meantime, you can use the Flutter SDK for cross-platform Android development.

Planned Features

  • Native Kotlin API with Java interop
  • Jetpack Compose integration
  • NNAPI acceleration
  • Vulkan GPU support
  • Qualcomm QNN support (Hexagon DSP)

Expected API

Based on the Flutter SDK patterns, the Android SDK will follow a similar structure:

import ai.xybrid.sdk.*

// Initialize
Xybrid.initialize(context)

// Load model from registry
val model = Xybrid.model(
    modelId = "whisper-tiny-candle",
    version = "1.0",
    registry = "https://registry.xybrid.dev"
).load()

// Run inference
val envelope = Envelope.audio(audioBytes)
val result = model.run(envelope)
println("Transcription: ${result.text}")

// Cleanup
model.unload()

Pipeline Example

val pipeline = Xybrid.pipeline(yaml = """
name: voice-assistant
stages:
  - whisper-tiny-candle@1.0
  - target: integration
    provider: openai
    model: gpt-4o-mini
  - kokoro-82m@0.1
""".trimIndent()).load()

val result = pipeline.run(
    envelope = Envelope.audio(audioBytes)
)

// Play TTS response
val audioOutput = result.audio

Stay Updated

Follow the GitHub repository for Android SDK announcements.

On this page