JVector - The most advanced embedded vector search engine

JVector is a pure Java, zero dependency, embedded vector search engine, used by DataStax Astra DB.

JVector uses state of the art graph algorithms inspired by DiskANN and related research that offer high recall and low latency. It uses the Panama SIMD API to accelerate index build and queries. It is memory efficient, compresses vectors using product quantization so they can stay in memory during searches.

JVector is concurrent, its index can scale linearly to 32 threads.