[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"glossary-vector-index::en":3,"gloss-cluster-vector-index::en":23,"gloss-next-vector-index::en":9},{"slug":4,"category":5,"name":6,"definition":7,"meta_desc":8,"faq":9,"schema_markup":9,"related":10},"vector-index","data-infra","Vector Index","A vector index is the data structure that makes similarity search fast. Comparing a query against every stored embedding is exact and linear; an index (HNSW, IVF, ScaNN and friends) narrows the search to a promising neighbourhood and answers in milliseconds instead of seconds.\n\nThe trade is approximation. These are approximate nearest-neighbour structures: they return the right answers most of the time, and the tuning parameters decide how often. Higher recall costs memory and latency; lower recall silently drops the one relevant document from a retrieval set nobody inspects.\n\nThat silence is why recall belongs in your monitoring. A RAG assistant with a badly tuned index does not error — it answers confidently from the four documents it did find, and the missing fifth is the one that contained the exception.\n\nPractical notes: filtered search (\"only this tenant's documents\") is where naive indexes fall over, so check that your database applies filters during traversal rather than after it. And rebuild cost matters — an index that takes six hours to rebuild shapes how often you can re-embed a corpus.","A vector index makes similarity search fast by trading exactness for speed — the recall you lose is a tuning decision, not a bug.",null,[11,14,17,20],{"slug":12,"name":13},"embedding","Embedding",{"slug":15,"name":16},"latency","Latency",{"slug":18,"name":19},"semantic-search","Semantic Search",{"slug":21,"name":22},"vector-database","Vector Database",[24,28,31,34,37,41,44,47,50,53,57,60],{"slug":25,"category":5,"name":26,"updated_at":27},"acid","ACID","2026-08-24T02:46:37+00:00",{"slug":29,"category":5,"name":30,"updated_at":27},"ann-search","ANN Search",{"slug":32,"category":5,"name":33,"updated_at":27},"backpressure","Backpressure",{"slug":35,"category":5,"name":36,"updated_at":27},"batch-processing","Batch Processing",{"slug":38,"category":5,"name":39,"updated_at":40},"bm25","BM25","2026-08-24T02:46:38+00:00",{"slug":42,"category":5,"name":43,"updated_at":27},"cache","Cache",{"slug":45,"category":5,"name":46,"updated_at":27},"cap-theorem","CAP Theorem",{"slug":48,"category":5,"name":49,"updated_at":27},"change-data-capture","Change Data Capture (CDC)",{"slug":51,"category":5,"name":52,"updated_at":27},"chroma","Chroma",{"slug":54,"category":5,"name":55,"updated_at":56},"chunk-overlap","Chunk Overlap","2026-08-24T03:30:02+00:00",{"slug":58,"category":5,"name":59,"updated_at":27},"columnar-storage","Columnar Storage",{"slug":61,"category":5,"name":62,"updated_at":27},"connection-pooling","Connection Pooling"]