Skip to main content
An Index in R2R represents a vector index structure designed to optimize similarity search operations across document chunks. Indices are a core component of R2R’s Retrieval-Augmented Generation (RAG) architecture, enabling fast and scalable semantic retrieval. By organizing vector embeddings efficiently, indices make it possible to perform high-performance vector searches across large datasets using different similarity metrics and indexing strategies.

​
Key Capabilities

Indices in R2R provide:
  • Fast similarity search for vector-based retrieval.
  • Multiple index methods, including HNSW (Hierarchical Navigable Small World) and IVF-Flat.
  • Configurable similarity measures such as cosine similarity or inner product.
  • Concurrent index building to improve throughput and scalability.
  • Performance optimization for large-scale vector operations.

​
API Endpoints

MethodEndpointDescription
GET/indicesList available indices with pagination.
GET/indices//Retrieve details and configuration for a specific index.
DELETE/indices//Delete an existing index.