# vchord Documentation

> Scalable, fast, and disk-friendly vector similarity search in PostgreSQL powered by RaBitQ quantization.

---

LLMS index: [llms.txt](/llms.txt)

---

**`vchord`** (VectorChord) is a PostgreSQL extension built on [`pgvector`](/docs/vector/) for scalable, high-performance, and disk-friendly vector similarity search. Powered by **RaBitQ** 1-bit/4-bit/8-bit quantization, it delivers inverted-file (`vchordrq`) and single-layer graph (`vchordg`) index access methods, native multi-vector MaxSim (`@#`), sphere range queries (`<<->>`), and external centroid precomputation.

## Documentation Sections


- [Getting Started](/docs/vchord/getting-started/) — Overview, quickstart tutorial, installation via Pigsty, Docker, or prebuilt packages, and migration from [`pgvecto.rs`](/docs/vectors/).
- [Usage Guide](/docs/vchord/usage/) — Indexing with `vchordrq` and `vchordg`, nearest-neighbor and sphere search, filtering, prewarm, external index builds, and full SQL reference.
- [Use Cases](/docs/vchord/use-case/) — Multi-vector ColBERT MaxSim (`@#`) late interaction and hybrid semantic + BM25 search with [`vchord_bm25`](/docs/bm25/) and [`pg_tokenizer`](/docs/tokenizer/).
- [Benchmarks \& TCO](/docs/vchord/benchmark/) — Reproducible performance, recall, memory, and cost comparisons across GIST-1M, Cohere-10M, and LAION-100M.
- [Administration](/docs/vchord/admin/) — Production parameter tuning, memory sizing, I/O concurrency, vacuuming, and version upgrades.
- [Customer Stories](/docs/vchord/customer-stories/) — How the University of Oxford stores and queries 3 billion satellite embedding vectors in PostgreSQL with `vchord`.


## Key Topics & Deep Dives


- [vchordrq vs vchordg Indexing](/docs/vchord/usage/indexing/) — Compare disk-friendly hierarchical K-Means (`vchordrq`) against low-latency single-layer graphs (`vchordg`) and configure build options.
- [RaBitQ4 \& RaBitQ8 Types](/docs/vchord/usage/quantization-types/) — Compress 32-bit vectors into 4-bit (`rabitq4`) or 8-bit (`rabitq8`) quantized types with theoretical error bounds.
- [Multi\-Vector MaxSim \(\@\#\)](/docs/vchord/usage/multi-vector-retrieval/) — Index `vector[]` token embeddings with WARP dynamic pruning for ColBERTv2 and ColPali visual document retrieval.
- [External Index Precomputation](/docs/vchord/usage/external-build/) — Offload K-Means centroid clustering to external CPUs or GPUs and build 100M-vector indexes in 20 minutes.
- [Performance Tuning \& Sizing](/docs/vchord/usage/postgresql-tuning/) — Tune `probes`, `epsilon`, `ef_search`, `io_search`, `max_scan_tuples`, and PostgreSQL shared memory for production workloads.
- [Migrating from pgvecto\.rs](/docs/vchord/admin/migration/) — Step-by-step SQL migration guide from [`pgvecto.rs`](/docs/vectors/) (`vectors` extension) or `pgvector` HNSW/IVFFlat to `vchord`.

---

Section pages:

- [Getting Started](/docs/vchord/getting-started/): Overview, installation packages, Docker images, and the Vchord retrieval suite.
- [Usage](/docs/vchord/usage/): Indexing with vchordrq and vchordg, MaxSim multi-vector retrieval, quantization types, range queries, and tuning.
- [Use Cases](/docs/vchord/use-case/): End-to-end recipes for hybrid BM25 + dense search, ColBERT reranking, and multimodal OCR-free RAG.
- [Benchmarks](/docs/vchord/benchmark/): Price, throughput, latency, and recall benchmarks against pgvector, pgvectorscale, and Elasticsearch.
- [Administration](/docs/vchord/admin/): Migrating from pgvecto.rs or pgvector, horizontal read scalability, and Kubernetes deployment.
- [Customer Stories](/docs/vchord/customer-stories/): Real-world production deployments of VectorChord at multi-billion vector scale.

---

Backlinks:

- [3 Billion Vectors in PostgreSQL to Protect the Earth](/blog/post/3-billion-vectors-in-postgresql-to-protect-the-earth/)
- [All-in-one VectorChord Suite: Building Production-Ready RAG Solutions Directly in PostgreSQL](/blog/post/all-in-one-vectorchord-suite-building-production-ready-rag-solutions-directly-in-postgresql/)
- [Hybrid search with Postgres Native BM25 and VectorChord](/blog/post/hybrid-search-with-postgres-native-bm25-and-vectorchord/)
- [Introducing Vechord: Turn PostgreSQL into your search engine in a Pythonic way.](/blog/post/introducing-vechord-turn-postgresql-into-your-search-engine-in-a-pythonic-way/)
- [Supercharge vector search with ColBERT rerank in PostgreSQL](/blog/post/supercharge-vector-search-with-colbert-rerank-in-postgresql/)
- [Vector Search Over PostgreSQL: A Comparative Analysis of Memory and Disk Solutions](/blog/post/vector-search-over-postgresql-a-comparative-analysis-of-memory-and-disk-solutions/)
- [VectorChord-BM25: Revolutionize PostgreSQL Search with BM25 Ranking — 3x Faster Than ElasticSearch](/blog/post/vectorchord-bm25-revolutionize-postgresql-search-with-bm25-ranking-3x-faster-than-elasticsearch/)
- [VectorChord: Cost-Efficient Upload & Search of 400 Million Vectors on AWS](/blog/post/vectorchord-cost-efficient-upload-and-search-of-400-million-vectors-on-aws/)
- [VectorChord: Store 400k Vectors for $1 in PostgreSQL](/blog/post/vectorchord-store-400k-vectors-for-1-in-postgresql/)
- [About](/docs/about/)
- [vchord_bm25](/docs/bm25/)
- [pgvector](/docs/vector/)
- [Getting Started](/docs/vector/getting-started/)
- [HNSW & IVFFlat Indexing](/docs/vector/indexing/)
- [Quantization & Sparse Vectors](/docs/vector/quantization/)
- [Querying & Filtering](/docs/vector/querying/)
- [Reference, Tuning & FAQ](/docs/vector/reference/)
- [pgvecto.rs](/docs/vectors/)