Getting Started — Android

speech-android provides on-device speech processing for Android using ONNX Runtime. The pipeline runs VAD + STT + TTS with barge-in support, fully offline after model download.

Requirements

Android 8+ (API 26)
arm64-v8a architecture

Download the pre-built demo app to try it immediately:

app-release.apk

Gradle Dependency

Add the SDK to your build.gradle.kts:

implementation("audio.soniqo:speech:0.0.5")

Quick Start

val modelDir = ModelManager.ensureModels(context)
val pipeline = SpeechPipeline(SpeechConfig(modelDir = modelDir))
pipeline.events.collect { event ->
    when (event) {
        is SpeechEvent.TranscriptionCompleted -> println(event.text)
        is SpeechEvent.ResponseDone -> pipeline.resumeListening()
        else -> {}
    }
}
pipeline.start()
pipeline.pushAudio(samples) // 16kHz mono float32

Important

Models auto-download from HuggingFace on first use (~1.2 GB total). After the initial download, all inference runs fully offline.

Models

All models run via ONNX Runtime with NNAPI acceleration. INT8 quantized by default.

Model	Task	Size
Parakeet TDT v3 (INT8)	Speech-to-Text (114 languages)	490 MB
Kokoro-82M (INT8)	Text-to-Speech (7 languages)	89 MB
Silero VAD v5	Voice Activity Detection	1.2 MB
DeepFilterNet3 (FP16)	Noise Cancellation	4.2 MB

Source code: github.com/soniqo/speech-android

Next Steps

Benchmarks — Android inference performance
Linux C API — embedded Linux setup