For AI agents: The documentation index is at https://docs.digitalocean.com/llms.txt. Markdown versions of pages use the same URL with index.html.md in place of the HTML page (for example, append index.html.md to the directory path instead of opening the HTML document).

Getting Started

Quickstart guides for using DigitalOcean Managed PostgreSQL with pgvector as a vector database.

How-Tos

How to enable pgvector, load embeddings, index and tune vector search, and build advanced workloads with pgvectorscale.

Concepts

Core PostgreSQL vector search concepts, including pgvector, pgvectorscale, embeddings, vector indexes, distance operators, and hybrid search.

DigitalOcean Managed PostgreSQL for vector search is the same managed PostgreSQL engine available under Managed Databases, with support for vector workloads through PostgreSQL extensions.

Use PostgreSQL for vector search when you want to store embeddings alongside relational data, query vectors with SQL, and keep one database as your source of truth.

PostgreSQL vector databases support:

• vector: The pgvector extension for vector, halfvec, and sparsevec columns, exact search, HNSW indexes, and IVFFlat indexes. Available on PostgreSQL 13 and later.
• vectorscale: The pgvectorscale extension for StreamingDiskANN and Statistical Binary Quantization for larger, disk-resident vector workloads. Available on PostgreSQL 14 and later.

Enable each extension per database with CREATE EXTENSION.

When to Choose PostgreSQL

Choose PostgreSQL for vector workloads when you need:

• Vectors with relational data: Combine vector similarity search with SQL filters, joins, and transactions.
• One database: Store source data, metadata, and embeddings together without syncing to a separate vector store.
• PostgreSQL tooling: Use existing migrations, backups, replication, observability, and access-control patterns.
• Small to medium vector datasets: Use pgvector HNSW or IVFFlat indexes for common vector search workloads.

For larger search-focused workloads, consider OpenSearch or Weaviate. For hybrid keyword and vector search as a first-class search workflow, consider OpenSearch.

What Is Included

PostgreSQL vector databases include the following vector search capabilities:

• pgvector vector extension: Available on PostgreSQL 13 and later.
• pgvectorscale vectorscale extension: Available on PostgreSQL 14 and later.
• HNSW and IVFFlat indexes: Supported through pgvector.
• StreamingDiskANN and SBQ: Supported through vectorscale.
• Hybrid full-text and vector search: Supported with tsvector, GIN indexes, and vector indexes.
• In-database embedding generation: Not supported. Generate embeddings in your application or ingest pipeline, then insert them into PostgreSQL.

PostgreSQL doesn’t generate embeddings. Generate embeddings in your application or ingest pipeline, then insert the resulting vectors into PostgreSQL.

Use Existing PostgreSQL Clusters

PostgreSQL clusters created through Vector Databases and Managed Databases use the same managed PostgreSQL engine. If you already have a compatible Managed PostgreSQL cluster, enable vector search with:

CREATE EXTENSION vector;

Latest Updates

3 August 2026

• PostgreSQL Advanced Edition clusters now support sslmode=verify-full for full TLS certificate verification. For setup, see How to Connect to PostgreSQL Database Clusters and Increase TLS Verification with sslmode.

15 July 2026

• Managed databases are available in Kansas City (mkc1) with limited plan availability. Premium AMD plans and General Purpose MongoDB clusters are not available in this region. See the regional availability matrix for current coverage by engine.

10 July 2026

• PostgreSQL Standard Edition clusters now support sslmode=verify-full for full TLS certificate verification. For setup, see How to Connect to PostgreSQL Database Clusters and Increase TLS Verification with sslmode.

For more information, see the full release notes.