Aura Core

Ultra-fast offline pattern classification engine built in Rust. Encodes text into Sparse Distributed Representations (SDR) — 256K-bit sparse binary vectors — and performs similarity-based retrieval using Tanimoto coefficient matching with inverted bitmap indexing. Achieves 60-200 microsecond classification in a 3MB binary with zero cloud dependencies. Designed for sovereign edge deployment: IoT devices, defense platforms, privacy-sensitive applications. Features include append-only crash-safe storage, RoaringBitmap O(k) indexing, full Unicode support, and 25-40% noise tolerance. 146 Rust tests, 212 Python integration tests. Patent Pending.