dba-vectordb

Name: dba-vectordb
Availability: InStock
Author: massimilianopili

Design pgvector schemas and HNSW/IVFFlat indexes for hybrid search.

Updated Mar 7, 2026

Search Tags:#semantic search #vector database #RAG #hybrid search #pgvector #HNSW #embedding

Ready to run

npx skills add https://github.com/MassimilianoPili/agent-framework --skill dba-vectordb

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: dba-vectordb
Source: https://github.com/MassimilianoPili/agent-framework/tree/main/.claude/agents/dba-vectordb
Command: npx skills add https://github.com/MassimilianoPili/agent-framework --skill dba-vectordb

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill streamlines the design and optimization of vector databases, crucial for efficient semantic search and Retrieval Augmented Generation (RAG) in AI applications.

Core Features & Use Cases

Vector Indexing: Configure HNSW and IVFFlat indexes for pgvector.
Hybrid Search: Implement combined vector similarity and BM25 search.
Embedding Pipelines: Design and optimize the ingestion of embedding data.
Use Case: Optimize a pgvector database for a RAG system by configuring an HNSW index for fast semantic search and implementing hybrid search to combine vector results with keyword relevance.

Quick Start

Use the dba-vectordb skill to create a pgvector HNSW index for embeddings in the 'documents' table.

Frequently Asked Questions about dba-vectordb

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I configure a pgvector HNSW index for semantic search?▼

To configure a pgvector HNSW index for semantic search, you tune indexing parameters to balance recall and speed. This involves designing the vector database schema and optimizing the HNSW graph configuration specifically for your RAG application query patterns.

What is hybrid search and how does it combine vector similarity with BM25?▼

Hybrid search combines vector similarity with BM25 keyword relevance to improve retrieval accuracy. It merges semantic embedding results with traditional text matching, ensuring RAG applications capture both contextual meaning and exact keyword matches.

HNSW vs IVFFlat: which pgvector index strategy should I use?▼

Choosing between HNSW and IVFFlat pgvector indexes depends on your recall and speed trade-offs. HNSW offers faster query speeds with higher memory usage, while IVFFlat provides faster build times and lower memory overhead for your vector database.

How do I design an embedding pipeline for RAG ingestion?▼

Designing an embedding pipeline for RAG ingestion involves structuring data transformation and vector embedding workflows. You optimize the pipeline to balance recall and ingestion speed, ensuring the vector database receives efficiently formatted semantic embeddings.

When should I not use HNSW indexes for my vector database?▼

You should avoid HNSW indexes when memory capacity is strictly limited or index build time is a critical bottleneck. In these constraint scenarios, IVFFlat serves as a more resource-efficient alternative for your pgvector semantic search implementation.