HelixDB

HelixDB is an OLTP graph database built in Rust on object storage. It aims to hold entities, relationships, embeddings and text together in a single transactional engine, so that applications such as retrieval-augmented generation can query graph structure, vectors and keywords from one place rather than stitching several databases together.

Data is modeled as a labeled property graph with typed properties. On top of that it provides approximate nearest-neighbor vector search that can be prefiltered by graph traversal, and BM25 keyword search over node and edge properties. You can query it from Rust, TypeScript, Go or Python. A command-line tool creates projects and starts a local instance, which needs Docker or Podman; data lives in memory by default and a disk option persists it across restarts.

HelixDB is licensed under Apache-2.0. You can run it yourself locally or on your own infrastructure. It suits developers building AI applications, knowledge graphs and recommendation features who want graph and vector retrieval together.