pgvector
A PostgreSQL extension that adds vector similarity search capabilities, enabling AI applications on existing PostgreSQL infrastructure.
In-Depth Explanation
pgvector is a PostgreSQL extension that adds vector data types and similarity search operators. It enables storing embeddings and performing vector search directly in PostgreSQL, leveraging existing database infrastructure.
Key features:
- PostgreSQL native: Standard SQL
- Vector type: Store embeddings
- Operators: Cosine, L2, inner product
- Indexing: IVFFlat, HNSW
- Hybrid queries: Combine with SQL
- Transactions: ACID compliance
Index types:
- IVFFlat: Fast for updates
- HNSW: Fast for queries
Use cases:
- Semantic search in existing apps
- RAG with PostgreSQL backend
- Recommendation systems
- Similarity matching
Business Context
pgvector is ideal for organisations already using PostgreSQL, adding AI capabilities without new infrastructure or data movement.
How Clever Ops Uses This
We implement pgvector for Australian businesses with existing PostgreSQL deployments, enabling AI features without additional database complexity.
Example Use Case
"Adding semantic search to an existing e-commerce PostgreSQL database: add vector column for product embeddings, enable similarity queries alongside existing SQL."
Frequently Asked Questions
Related Terms
Related Resources
Vector Database
A specialised database optimised for storing and searching vector embeddings. Es...
Embeddings
Numerical vector representations of text, images, or other data that capture sem...
Learning Centre
Guides, articles, and resources on AI and automation.
AI & Automation Services
Explore our full AI automation service offering.
AI Readiness Assessment
Check if your business is ready for AI automation.