expo-ai-kit
A TypeScript library that adds on-device text generation, speech-to-text and vision to Expo and React Native apps, using each platform's native AI models with no API keys or cloud calls.
expo-ai-kit brings on-device AI to Expo and React Native apps: local language model text generation, speech-to-text, computer vision and embeddings, all without an API key or a cloud service. Every feature is optional, so unused native code, models and permissions are left out of the build.
One TypeScript API covers both platforms: Apple Foundation Models, Speech and Vision on iOS, and ML Kit and MediaPipe on Android, with a fallback runtime for downloadable models such as Gemma, Qwen and Phi. Prompts, audio and images never leave the device.
It requires Expo SDK 54 or newer, or compatible Expo modules in an existing React Native app, and a native build; the standard Expo Go client is not supported.