AI & Automation
What is Vector Database?
Definition
A database optimised for storing and searching high-dimensional numerical representations of meaning (embeddings) — the storage layer that makes semantic search and RAG possible.
In more detail
A vector database stores data not as rows and columns, but as vectors — lists of numbers that represent the semantic meaning of text, images, or other content. These numerical representations (called embeddings) are generated by AI models and capture meaning in a mathematical space where similar concepts are numerically close to each other.
This enables semantic search: instead of searching for exact keyword matches, you search for meaning. Ask 'what are the payment terms?' and a vector search will find the relevant contract clause even if it never uses those exact words. This is the retrieval mechanism that powers RAG systems.
Popular vector database options include Pinecone (managed, cloud-native), Weaviate (open-source), Qdrant (high-performance), and pgvector (a PostgreSQL extension that adds vector capabilities to a standard relational database). For many projects, pgvector is the practical choice because it avoids adding a new infrastructure component.
Why it matters
If you're building any AI system that needs to work with your own documents, knowledge base, or data — whether that's a Q&A system, a semantic search feature, or an agentic workflow that retrieves context — a vector database is likely part of the infrastructure.
Related terms
Related service
Working with Vector?
I offer AI Integration & Agentic Workflows for businesses ready to move from understanding to implementation.
Learn about AI Integration & Agentic Workflows →