VortexDB
A vector database built from scratch in Rust, organized as a workspace of modular crates with pluggable indexers, storage backends, and HTTP, gRPC and terminal clients.
VortexDB is a vector database written in Rust as a modular workspace of crates, designed for similarity search and approximate nearest-neighbor search behind AI-powered applications.
Its indexing layer supports Flat brute-force search, KD-Tree space partitioning, and HNSW graphs for large-scale approximate search, with Euclidean, Manhattan, Hamming and Cosine distance metrics. Storage is abstracted behind a StorageEngine trait, with RocksDB for persistent production storage and an in-memory backend for tests. It ships an Axum-based HTTP REST server, a gRPC server with full CRUD for vector points, and a Ratatui-based terminal client for browsing and managing the database, plus a Docker Compose setup for running the HTTP and gRPC servers together.
It is licensed under Apache-2.0, and its documentation lives at vortex-db.sdslabs.co.