React Native RAG
A modular React Native library for building private, local retrieval-augmented-generation apps, with pluggable LLM, embeddings, vector-store and text-splitter components and on-device inference through ExecuTorch.
React Native RAG offers three ways to add RAG to a React Native app: a simple useRAG hook for quick setups, a more configurable RAG class, or direct use of its individual LLM, Embeddings, VectorStore and TextSplitter components for fine-grained control. Each interface is implemented against documented component interfaces, so a project can swap in custom implementations of any piece.
On-device inference is provided by a companion @react-native-rag/executorch package for private, efficient model execution directly on the user's device, and @react-native-rag/op-sqlite persists vector stores locally with SQLite; the VectorStore and Embeddings components can also be used on their own to add semantic search to an app. The library powers Private Mind, a privacy-first mobile AI app on the App Store and Google Play, as a real-world example of its use.