datastax/jvector
JVector is an embedded Java vector search engine implementing HNSW and DiskANN graph-based ANN algorithms for high-dimensional similarity search.
Not currently ranked — collecting fresh signals.
star history
JVector is an embedded vector search engine for approximate nearest neighbor (ANN) search in high-dimensional spaces. It implements a multi-layer graph index that merges the HNSW and DiskANN/Vamana algorithm designs, featuring nonblocking concurrency control for linear scalability across CPU cores. The engine addresses the curse of dimensionality by providing logarithmic-time approximate queries rather than exact but linear-time results.
Frequently asked
- What is datastax/jvector?
- JVector is an embedded Java vector search engine implementing HNSW and DiskANN graph-based ANN algorithms for high-dimensional similarity search.
- Is jvector open source?
- Yes — datastax/jvector is open source, released under the Apache-2.0 license.
- What language is jvector written in?
- datastax/jvector is primarily written in Java.
- How popular is jvector?
- datastax/jvector has 1.7k stars on GitHub.
- Where can I find jvector?
- datastax/jvector is on GitHub at https://github.com/datastax/jvector.