# LynseDB

A Python-first vector database with a Rust storage and search core, designed to move from an embedded local prototype to an HTTP service and then a lightweight sharded cluster without changing the client API.

LynseDB is a vector database with a Python client and a Rust storage and search engine, aimed at retrieval-augmented generation, semantic search, agent memory, multimodal retrieval, document question answering and recommendation features.

The same client API works whether the database runs embedded inside a single Python process, as a single HTTP server, or as a coordinator-backed cluster with shard groups, hash-based routing, replica write mirroring and health checks. Beyond embedding search, it exposes native distance functions for geospatial coordinates, packed binary fingerprints, aligned numeric profiles and probability distributions, alongside dense vectors, metadata filters, BM25, sparse vectors, named vector fields and external reranking through one collection API.

Production deployments can add API keys, health and readiness checks, metrics, OpenAPI, Docker, systemd and Kubernetes examples on top of the same core. It is licensed under Apache-2.0.

Website: https://birchkwok.github.io/lynsedb