VectorIndexKind

Enums · AiDotNet.Enums

Selects the in-memory vector-search index used to back a RAG document store.

For Beginners

A vector index is the data structure that makes “find the most similar vectors” fast. Different indexes trade accuracy for speed:

  • Flat - exact brute-force search; most accurate, slowest on large sets.
  • HNSW - graph-based approximate search; excellent speed/recall balance.
  • IVF - inverted-file clustering; fast on large collections.
  • LSH - locality-sensitive hashing; very fast, lower recall.

Fields

Field Summary
VectorIndexKind Flat Exact brute-force flat index (most accurate).
VectorIndexKind HNSW Hierarchical Navigable Small World graph index (fast, high recall).
VectorIndexKind IVF Inverted File index with clustering (scales to large collections).
VectorIndexKind LSH Locality-Sensitive Hashing index (very fast, approximate).
int value__