Login
Login

vector database

{{Short description|Type of specialized database system}}

{{Machine learning}}

A vector database, vector store or vector search engine is a database that can store vectors (fixed-length lists of numbers) along with other data items. Vector databases typically implement one or more Approximate Nearest Neighbor algorithms,{{cite web |author1=Roie Schwaber-Cohen |title=What is a Vector Database & How Does it Work |url=https://www.pinecone.io/learn/vector-database/ |access-date=18 November 2023 |publisher=Pinecone}}{{cite web |title=What is a vector database |url=https://www.elastic.co/what-is/vector-database |access-date=18 November 2023 |publisher=Elastic}}{{cite web |title=What is a Vector Database? |url=https://www.datastax.com/guides/what-is-a-vector-database |access-date=10 July 2023}} so that one can search the database with a query vector to retrieve the closest matching database records.

Vectors are mathematical representations of data in a high-dimensional space. In this space, each dimension corresponds to a feature of the data, with the number of dimensions ranging from a few hundred to tens of thousands, depending on the complexity of the data being represented. A vector's position in this space represents its characteristics. Words, phrases, or entire documents, as well as images, audio, and other types of data, can all be vectorized.{{Cite web |last= |date=2023-12-26 |title=Vector database |url=https://learn.microsoft.com/en-us/azure/cosmos-db/vector-database |access-date=2024-01-11 |website=learn.microsoft.com |language=}}

These feature vectors may be computed from the raw data using machine learning methods such as feature extraction algorithms, word embeddings{{Cite web |author=Evan Chaki |date=2023-07-31 |title=What is a vector database? |url=https://learn.microsoft.com/en-us/semantic-kernel/memories/vector-db |accessdate= |publisher=Microsoft |quote=A vector database is a type of database that stores data as high-dimensional vectors, which are mathematical representations of features or attributes.}} or deep learning networks. The goal is that semantically similar data items receive feature vectors close to each other.

Vector databases can be used for similarity search, semantic search, multi-modal search, recommendations engines, large language models (LLMs), object detection, etc.{{Cite web |last= |date=2023-12-26 |title=Vector database |url=https://learn.microsoft.com/en-us/azure/cosmos-db/vector-database |access-date=2024-01-11 |website=learn.microsoft.com |language=}}

Vector databases are also often used to implement retrieval-augmented generation (RAG), a method to improve domain-specific responses of large language models. The retrieval component of a RAG can be any search system, but is most often implemented as a vector database. Text documents describing the domain of interest are collected, and for each document or document section, a feature vector (known as an "embedding") is computed, typically using a deep learning network, and stored in a vector database. Given a user prompt, the feature vector of the prompt is computed, and the database is queried to retrieve the most relevant documents. These are then automatically added into the context window of the large language model, and the large language model proceeds to create a response to the prompt given this context.{{cite journal |last1=Lewis |first1=Patrick |last2=Perez |first2=Ethan |last3=Piktus |first3=Aleksandra |last4=Petroni |first4=Fabio |last5=Karpukhin |first5=Vladimir |last6=Goyal |first6=Naman |last7=Küttler |first7=Heinrich |title=Retrieval-augmented generation for knowledge-intensive NLP tasks |journal=Advances in Neural Information Processing Systems 33 |year=2020 |pages=9459–9474|arxiv=2005.11401 }}

Techniques

The most important techniques for similarity search on high-dimensional vectors include:

and combinations of these techniques.{{Citation needed|date=March 2024}}

In recent benchmarks, HNSW-based implementations have been among the best performers.{{Citation |last1=Aumüller |first1=Martin |title=ANN-Benchmarks: A Benchmarking Tool for Approximate Nearest Neighbor Algorithms |date=2017 |work=Similarity Search and Applications |volume=10609 |pages=34–49 |editor-last=Beecks |editor-first=Christian |url=http://link.springer.com/10.1007/978-3-319-68474-1_3 |access-date=2024-03-19 |place=Cham |publisher=Springer International Publishing |doi=10.1007/978-3-319-68474-1_3 |isbn=978-3-319-68473-4 |last2=Bernhardsson |first2=Erik |last3=Faithfull |first3=Alexander |editor2-last=Borutta |editor2-first=Felix |editor3-last=Kröger |editor3-first=Peer |editor4-last=Seidl |editor4-first=Thomas|arxiv=1807.05614 }}{{Cite book |last1=Aumüller |first1=Martin |last2=Bernhardsson |first2=Erik |last3=Faithfull |first3=Alexander |chapter=ANN-Benchmarks: A Benchmarking Tool for Approximate Nearest Neighbor Algorithms |series=Lecture Notes in Computer Science |date=2017 |volume=10609 |editor-last=Beecks |editor-first=Christian |editor2-last=Borutta |editor2-first=Felix |editor3-last=Kröger |editor3-first=Peer |editor4-last=Seidl |editor4-first=Thomas |title=Similarity Search and Applications |chapter-url=https://link.springer.com/chapter/10.1007/978-3-319-68474-1_3 |language=en |location=Cham |publisher=Springer International Publishing |pages=34–49 |doi=10.1007/978-3-319-68474-1_3 |isbn=978-3-319-68474-1|arxiv=1807.05614 }} Conferences such as the International Conference on Similarity Search and Applications, SISAP and the Conference on Neural Information Processing Systems (NeurIPS) host competitions on vector search in large databases.

Implementations

{{Dynamic list|multiple=no}}

class="wikitable sortable "

|+

NameLicense
Aerospike{{Cite web |date=2024-05-07 |title=Aerospike Recognized by Independent Research Firm Among Notable Vendors in Vector Databases Report

|url=https://www.morningstar.com/news/globe-newswire/9111790/aerospike-recognized-by-independent-research-firm-among-notable-vendors-in-vector-databases-report |access-date=2024-08-01 |website=Morningstar |language=en-US}}{{Cite web |date=2024-04-04 |title=Aerospike raises $109M for its real-time database platform to capitalize on the AI boom

|url=https://techcrunch.com/2024/04/04/aerospike-raises-100m-for-its-real-time-database-platform-to-capitalize-on-the-ai-boom/ |access-date=2024-08-01 |website=TechCrunch |language=en-US}}

|Proprietary

AllegroGraph{{Cite web |date=2023-12-29 |title=AllegroGraph 8.0 Incorporates Neuro-Symbolic AI, a Pathway to AGI

|url=https://thenewstack.io/allegrograph-8-0-incorporates-neuro-symbolic-ai-a-pathway-to-agi/ |access-date=2024-06-06 |website=TheNewStack |language=en-US}}{{Cite web |date=2024-01-18 |title=Franz Inc. Introduces AllegroGraph Cloud: A Managed Service for Neuro-Symbolic AI Knowledge Graphs

|url=https://www.datanami.com/this-just-in/franz-inc-introduces-allegrograph-cloud-a-managed-service-for-neuro-symbolic-ai-knowledge-graphs/ |access-date=2024-06-06 |website=Datanami |language=en-US}}

|Proprietary (Managed Service)

Apache Cassandra{{Cite web |date=2023-09-22 |title=5 Hard Problems in Vector Search, and How Cassandra Solves Them

|url=https://thenewstack.io/5-hard-problems-in-vector-search-and-how-cassandra-solves-them/ |access-date=2023-09-22 |website=TheNewStack |language=en-US}}{{Cite web |title=Vector Search quickstart

|url=https://cassandra.apache.org/doc/latest/cassandra/vector-search/overview.html |access-date=2023-11-21 }}

|Apache License 2.0

Chroma{{Cite web |last=Palazzolo |first=Stephanie |title=Vector database Chroma scored $18 million in seed funding at a $75 million valuation. Here's why its technology is key to helping generative AI startups.

|url=https://www.businessinsider.com/vector-database-startup-chroma-raises-seed-funding-generative-artificial-intelligence-2023-4 |access-date=2023-11-16 |website=Business Insider |language=en-US}}{{Cite web |last=MSV |first=Janakiram |date=2023-07-28 |title=Exploring Chroma: The Open Source Vector Database for LLMs

|url=https://thenewstack.io/exploring-chroma-the-open-source-vector-database-for-llms/ |access-date=2023-11-16 |website=The New Stack |language=en-US}}

|Apache License 2.0{{cite web

|url=https://github.com/chroma-core/chroma/blob/main/LICENSE|title=chroma/LICENSE at main · chroma-core/chroma|website=GitHub|language=en}}

Azure Cosmos DB{{Cite web |date= 26 December 2023|title=Vector database

|url=https://learn.microsoft.com/azure/cosmos-db/vector-database |access-date=2024-01-10 |website=learn.microsoft.com}}

|Proprietary (Managed Service)

Couchbase{{Cite web |date=2023-08-30 |title=Couchbase aims to boost developer database productivity with Capella IQ AI tool

|url=https://venturebeat.com/ai/couchbase-aims-to-boost-developer-database-productivity-with-capella-iq-ai-tool/#h-next-on-the-roadmap-for-couchbase-is-vector-support|website=VentureBeat|language=en-US}}{{Cite web |date=2023-12-06 |title=Investor Presentation Third Quarter Fiscal 2024

|url=https://investors.couchbase.com/static-files/551e5b96-5307-4119-b225-19cfd8540242|website=Couchbase Investor Relations|language=en-US}}

|BSL 1.1{{Cite web |last=Anderson |first=Scott |date=2021-03-26 |title=Couchbase Adopts BSL License |url=https://www.couchbase.com/blog/couchbase-adopts-bsl-license/ |access-date=2024-02-14 |website=The Couchbase Blog |language=en-US}}

CrateDB{{Cite web |date= 16 November 2023|title=Open Source Vector Database

|url=https://cratedb.com/blog/open-source-vector-database |access-date=2024-11-06 |website=CrateDB Blog}}

|Apache License 2.0

DataStax{{cite news |last1=Sean Michael Kerner |title=DataStax brings vector database search to multicloud with Astra DB |url=https://venturebeat.com/data-infrastructure/datastax-brings-vector-database-search-to-multicloud-with-astra-db/ |agency=Venture Beat |date=18 July 2023}}

|Proprietary (Managed Service)

Elasticsearch{{cite news |last1=Kerner|first1=Sean |title=Elasticsearch Relevance Engine brings new vectors to generative AI

|url=https://venturebeat.com/ai/elasticsearch-relevance-engine-brings-new-vectors-to-generative-ai/ |access-date=18 November 2023 |work=VentureBeat |date=23 May 2023}}

|Server Side Public License, Elastic License{{cite web

|url=https://github.com/elastic/elasticsearch/blob/main/LICENSE.txt|title=elasticsearch/LICENSE.txt at main · elastic/elasticsearch|website=GitHub|language=en}}

HAKES{{cite web |title=HAKES {{!}} Efficient Data Search with Embedding Vectors at Scale |url=https://www.comp.nus.edu.sg/~dbsystem/hakes |access-date=8 March 2025 |language=en}}

|Apache License 2.0{{cite web |title=HAKES/LICENSE at main · nusdbsystem/HAKES |url=https://github.com/nusdbsystem/HAKES/blob/main/LICENSE |website=GitHub |access-date=8 March 2025 |language=en}}

HDF5 Query Indexing{{cite web |title=HDF5 Query Indexing

|website=GitHub |url=https://github.com/HDFGroup/hdf5doc/tree/master/RFCs/HDF5/Query-Indexing |access-date=3 May 2024 |date= 27 Sep 2019}}

|BSD 3-Clause{{Cite web |title=HDFGroup/COPYING at master · HDFGroup/hdf5

|url=https://github.com/HDFGroup/hdf5/blob/master/COPYING |access-date=2023-10-29 |website=GitHub |language=en}}

JaguarDB{{Cite web |title=JaguarDB Homepage

|url=http://jaguardb.com/ |access-date=2025-04-12 |website=JaguarDB|language=en-US}}{{cite web|date=2023-07-03|title=Vector DBMS| url=https://db-engines.com/de/ranking/vector+dbms | access-date=2025-04-12 |website=db-engines.com |language=en-US}}

|Proprietary

LanceDB{{Cite web |date=2024-12-17 |title=LanceDB Homepage

|url=https://lancedb.com/ |access-date=2024-12-17 |website=LanceDB |language=en-US}}

|Apache License 2.0{{Cite web |title=lancedb/LICENSE at main · lancedb/lancedb

|url=https://github.com/lancedb/lancedb?tab=Apache-2.0-1-ov-file |access-date=2024-12-17 |website=GitHub |language=en}}

Lantern{{Cite web |date=2024-04-05 |title=Lantern

|url=https://lantern.dev/ |access-date=2024-04-05 |language=en-US}}

|BSL 1.1{{Cite web |title=lantern/LICENSE at main /lanterndata/lantern |url=https://github.com/lanterndata/lantern/blob/main/LICENSE

|access-date=2024-04-10 |website=GitHub

|language=en}}

LlamaIndex{{Cite web |last=Wiggers |first=Kyle |date=2023-06-06 |title=LlamaIndex adds private data to large language models

|url=https://techcrunch.com/2023/06/06/llamaindex-adds-private-data-to-large-language-models/ |access-date=2023-10-29 |website=TechCrunch |language=en-US}}

|MIT License{{Cite web |title=llama_index/LICENSE at main · run-llama/llama_index

|url=https://github.com/run-llama/llama_index/blob/main/LICENSE |access-date=2023-10-29 |website=GitHub |language=en}}

MariaDB{{Cite web |title=MariaDB Vector |url=https://mariadb.org/projects/mariadb-vector/ |access-date=2024-07-30 |website=MariaDB.org |language=en-US}}{{Cite web |title=Vector search in old and modern databases |url=https://manticoresearch.com/blog/vector-search-in-databases/ |access-date=2024-07-30 |website=manticoresearch.com |language=en-us}}

|GPL v2{{Cite web |title=Licensing FAQ |url=https://mariadb.com/kb/en/licensing-faq/ |access-date=2024-07-30 |website=MariaDB KnowledgeBase}}

Marqo{{Cite web |last=Sawers |first=Paul |date=2023-08-16 |title=Meet Marqo, an open source vector search engine for AI applications |url=https://techcrunch.com/2023/08/16/meet-marqo-an-open-source-vector-search-engine-for-ai-applications/ |access-date=2024-08-20 |website=TechCrunch |language=en-US}}

|Apache License 2.0{{Citation |title=marqo-ai/marqo |date=2024-08-20 |url=https://github.com/marqo-ai/marqo?tab=Apache-2.0-1-ov-file#readme |access-date=2024-08-20 |publisher=Marqo}}

Meilisearch{{Cite web |date=2024-10-08 |title=Meilisearch Homepage

|url=https://meilisearch.com/ |access-date=2023-10-29 |website=Meilisearch |language=en-US}}

|MIT License{{Cite web |title=meilisearch/LICENSE at main · meilisearch/meilisearch

|url=https://github.com/meilisearch/meilisearch/blob/main/LICENSE |access-date=2024-10-08 |website=GitHub |language=en}}

Milvus{{cite web

|url=https://milvus.io/ |title=Open Source Vector Database – Milvus – LFAI & DATA |access-date=29 October 2023}}{{Cite web |last=Liao |first=Ingrid Lunden and Rita |date=2022-08-24 |title=Zilliz raises $60M, relocates to SF

|url=https://techcrunch.com/2022/08/24/zilliz-the-startup-behind-the-milvus-open-source-vector-database-for-ai-applications-raises-60m-and-relocates-to-sf/ |access-date=2023-10-29 |website=TechCrunch |language=en-US}}

|Apache License 2.0

MongoDB Atlas{{Cite web |date=2023-06-22 |title=Introducing Atlas Vector Search: Build Intelligent Applications with Semantic Search and AI Over Any Type of Data

|url=https://www.mongodb.com/blog/post/introducing-atlas-vector-search-build-intelligent-applications-semantic-search-ai|website=MongoDB|language=en-US}}

|Server Side Public License (Managed service)

Neo4j

{{Cite web |date=2023-08-22|url=https://itbrief.com.au/story/neo4j-enhances-its-graph-database-with-vector-search|website=itbrief|language=en-AU|title=Neo4j enhances its graph database with vector search}}

{{Cite web |url=https://neo4j.com/docs/cypher-manual/current/indexes/semantic-indexes/vector-indexes|website=neo4j|language=en-US|title=Vector search indexes

}}

|GPL v3 (Community Edition){{cite web

|url=https://neo4j.com/licensing/|title=Neo4j licensing}}

ObjectBox{{cite web|date=2024-07-03|title=Top Fifteen Vector Databases| url=https://db-engines.com/de/ranking/vektor+dbms | access-date=2024-07-03 |website=db-engines.com |language=en-US}}

|Apache License 2.0{{cite web

|url=https://github.com/objectbox/objectbox-java/blob/main/LICENSE.txt|website=github|title=ObjectBox Java license}}

OpenSearch{{cite web|date=2023-08-02|title=Using OpenSearch as a Vector Database| url=https://opensearch.org/platform/search/vector-database.html | access-date=2024-02-07 |website=OpenSearch.org |language=en-US}}{{Citation |last1=Pan |first1=James Jie |title=Survey of Vector Database Management Systems |date=2023-10-21 |arxiv=2310.14021 |last2=Wang |first2=Jianguo |last3=Li |first3=Guoliang}}{{Cite web |date=2023-07-26 |title=AWS debuts new AI-powered data management and analysis tools |url=https://siliconangle.com/2023/07/26/aws-debuts-new-ai-powered-data-management-analysis-tools/ |access-date=2024-02-07 |website=SiliconANGLE |language=en-US}}

|Apache License 2.0{{cite web

|url=https://github.com/opensearch-project/OpenSearch/blob/main/LICENSE.txt|website=github|title=OpenSearch license}}

Oracle Database{{Cite web |last=Hook(1) and Priyadarshi(2) |first=Doug(1) and Ranjan(2) |date=May 2, 2024 |title=Oracle Announces General Availability of AI Vector Search in Oracle Database 23ai |url=https://blogs.oracle.com/database/post/oracle-announces-general-availability-of-ai-vector-search-in-oracle-database-23ai |access-date=July 9, 2024 |website=oracle}}

|Proprietary (Managed Service or License)

Pinecone{{Cite web |date=2023-07-14 |title=Pinecone leads 'explosion' in vector databases for generative AI

|url=https://venturebeat.com/ai/pinecone-leads-explosion-in-vector-databases-for-generative-ai/ |access-date=2023-10-29 |website=VentureBeat |language=en-US}}

|Proprietary (Managed Service)

Postgres with pgvector{{Cite web |title=pgvector

|url=https://github.com/pgvector/pgvector |access-date=2023-11-27 |website=GitHub |language=en-US}}

| PostgreSQL License{{Cite web |title=pgvector/License

|url=https://github.com/pgvector/pgvector/blob/master/LICENSE |access-date=2023-11-27 |website=GitHub |language=en-US}}

Qdrant{{Cite web |last=Sawers |first=Paul |date=2023-04-19 |title=Qdrant, an open-source vector database startup, wants to help AI developers leverage unstructured data

|url=https://techcrunch.com/2023/04/19/qdrant-an-open-source-vector-database-startup-wants-to-help-ai-developers-leverage-unstructured-data/ |access-date=2023-10-29 |website=TechCrunch |language=en-US}}

|Apache License 2.0{{Cite web |title=qdrant/LICENSE at master · qdrant/qdrant

|url=https://github.com/qdrant/qdrant/blob/master/LICENSE |access-date=2023-10-29 |website=GitHub |language=en}}

Redis Stack{{Cite web |title=Using Redis as a Vector Database with OpenAI {{!}} OpenAI Cookbook |url=https://cookbook.openai.com/examples/vector_databases/redis/getting-started-with-redis-and-openai |access-date=2024-02-10 |website=cookbook.openai.com |language=en}}{{Cite web |title=Redis as a vector database quick start guide

|url=https://redis.io/docs/get-started/vector-database/ |access-date=2024-01-31 |website=Redis |language=en}}

|[https://redis.io/docs/about/license/ Redis Source Available License]{{Cite web |title=Search and query |url=https://redis.io/docs/interact/search-and-query/ |access-date=2024-02-10 |website=Redis |language=en}}

Snowflake{{Cite web |date=2024-05-17 |title=Vector data type and vector similarity functions — General Availability

|url=https://docs.snowflake.com/en/release-notes/2024/other/2024-05-16-vector-data-type-ga |access-date=2024-05-17 |website=Snowflake |language=en-US}}

|Proprietary (Managed Service)

SurrealDB{{Cite web |last=Wiggers |first=Kyle |date=2023-01-04 |title=SurrealDB raises $6M for its database-as-a-service offering

|url=https://techcrunch.com/2023/01/04/surrealdb-raises-6m-startup-funding-database-as-a-service/ |access-date=2024-01-19 |website=TechCrunch |language=en-US}}

|BSL 1.1{{Cite web |title=SurrealDB {{!}} License FAQs {{!}} The ultimate multi-model database |url=https://surrealdb.com/license |access-date=2024-02-14 |website=SurrealDB |language=en}}

Typesense{{Cite web |last=Martinez |first=Miguel |date=2024-06-20 |title=Typesense Homepage

|url=https://typesense.org/ |access-date=2024-06-20 |website=Typesense |language=en-US}}

|GPL v3 (Community Edition){{cite web

|url=https://github.com/typesense/typesense/blob/main/LICENSE.txt|title=Typesense licensing|website=GitHub }}

Vespa{{cite news |last1=Riley |first1=Duncan |title=Yahoo spins off AI scaling engine Vespa as an independent company

|url=https://siliconangle.com/2023/10/04/yahoo-spins-off-ai-scaling-engine-vespa-independent-company/ |access-date=18 November 2023 |work=siliconANGLE |date=4 October 2023}}

|Apache License 2.0{{cite web

|url=https://github.com/vespa-engine/vespa/blob/master/LICENSE|title=vespa/LICENSE at master · vespa-engine/vespa|website=GitHub|language=en}}

Weaviate{{Cite web |date=2023-04-21 |title=Weaviate reels in $50M for its AI-optimized vector database

|url=https://siliconangle.com/2023/04/21/weaviate-reels-50m-ai-optimized-vector-database/ |access-date=2023-10-29 |website=SiliconANGLE |language=en-US}}

|BSD 3-Clause{{Cite web |title=weaviate/LICENSE at master · weaviate/weaviate

|url=https://github.com/weaviate/weaviate/blob/master/LICENSE |access-date=2023-10-29 |website=GitHub |language=en}}

See also

  • {{annotated link|Curse of dimensionality}}
  • {{annotated link|Machine learning}}
  • {{annotated link|Nearest neighbor search}}
  • {{annotated link|Recommender system}}

References

{{Reflist}}

External links

  • {{cite web |last=Sawers |first=Paul |title=Why vector databases are having a moment as the AI hype cycle peaks |website=TechCrunch |date=2024-04-20 |url=https://techcrunch.com/2024/04/20/why-vector-databases-are-having-a-moment-as-the-ai-hype-cycle-peaks/ |access-date=2024-04-23}}

Category:Machine learning

Category:Types of databases