pgvector-semantic-search

What problem does it solve?

This skill enables storing, indexing, and querying high-dimensional embeddings inside PostgreSQL using pgvector, enabling scalable similarity search without leaving the database.

Core Features & Use Cases

• Vector storage & indexing: store embeddings as pgvector types and build ANN indexes (HNSW, IVFFlat) for fast retrieval.
• RAG integrations: support Retrieval-Augmented Generation pipelines by retrieving relevant vectors and documents for downstream tasks.
• Performance optimization: guidance on quantization, dimensionality considerations, memory tuning, and query patterns for large datasets.
• Use Case: Build a semantic search over documents or chat results with vector-based retrieval and ranking.

Quick Start

Install and configure the pgvector extension in your PostgreSQL database, create a vector column (halfvec or vector), and build an index (e.g., ON ... USING hnsw). Then run a nearest-neighbor query to retrieve similar items using a sample embedding.