# Aeru

An iOS app that answers questions by combining Apple's on-device FoundationModels with a local vector-search knowledge base and real-time web search, so responses can draw on both stored documents and current information.

Aeru is an iOS application that uses Apple's FoundationModels framework together with retrieval-augmented generation and web search to produce contextually aware answers. It searches a local vector database built with the NaturalLanguage framework for relevant stored context, and separately scrapes DuckDuckGo for up-to-date web content, streaming the model's response as it generates.

The app keeps conversation history for coherent multi-turn interactions, supports voice input for hands-free use, and can process documents in several formats into its knowledge base. It requires an Apple Intelligence-activated device running the latest iOS 26 and is built with SwiftUI, using SVDB for vector storage and SwiftSoup for parsing scraped web content.

Website: https://github.com/sskarz/Aeru