Term

Retrieval

The lookup stage of a RAG pipeline — fetching relevant text chunks from a corpus, given a query embedding. Quality of retrieval is usually the bottleneck on RAG quality, not the LLM itself.

Background

Retrieval strategies range from dense vector search to BM25 keyword search to hybrid combinations. Modern systems re-rank with a cross-encoder for precision. Recall-at-k and MRR are the standard metrics. Chunk size, overlap, and embedding model choice each shift the precision/recall curve meaningfully.

Tools that use it

01
Qdrant
Open-source, Rust-based vector database optimised for high-recall similarity search.
→
02
Weaviate
Open-source vector database with hybrid search and generative module integrations.
→
03
Pinecone
Managed vector database for RAG, semantic search, and recommendation at production scale.
→