React Native Sherpa-ONNX
Offline on-device speech processing for React Native. Add speech-to-text, text-to-speech, speaker diarization, and VAD to your mobile apps — completely offline, with full privacy.
Quick Start
Get up and running with Sherpa-ONNX in your React Native app in minutes.
Install the package
Add react-native-sherpa-onnx to your React Native project using your preferred package manager.For iOS, run
pod install to complete the setup. See the Installation guide for platform-specific details.Download a model
Download a pre-trained model for your use case. Models are available for speech-to-text, text-to-speech, and more.Browse available models in the Models section.
Initialize and use
Create a speech-to-text or text-to-speech engine and start processing audio.See the Quickstart for a complete example.
Core Features
Powerful speech processing capabilities for your React Native apps.
Speech-to-Text
Offline transcription with support for Whisper, Paraformer, Zipformer, and more.
Streaming STT
Real-time speech recognition with partial results and endpoint detection.
Text-to-Speech
High-quality voice synthesis with VITS, Matcha, Kokoro, and other TTS models.
Streaming TTS
Low-latency incremental speech generation for responsive user experiences.
Hardware Acceleration
Leverage NNAPI, XNNPACK, Core ML, and QNN for optimized performance.
Model Management
Flexible model loading from assets, file system, or Play Asset Delivery.
Explore by Topic
Dive deeper into specific features and integrations.
API Reference
Complete API documentation with types and examples.
Model Catalog
Browse supported STT and TTS models with download links.
Advanced Usage
PCM streaming, hotwords, and model detection.
Configuration
Optimize bundle size and configure platform features.
Migration Guide
Upgrade from previous versions with breaking changes.
Examples
Sample apps and real-world integration patterns.
Ready to get started?
Follow our quickstart guide to add offline speech processing to your React Native app in minutes.
Start Building