hai-vecdb

Name: hai-vecdb
Availability: InStock
Author: hai-series

Manage collections, vectors, and similarity search across LanceDB, pgvector, and Qdrant.

3|7|Updated Jan 28, 2026

Search Tags:#embeddings #rag #retrieval #vector-search #backends #vector-database #vecdb

Ready to run

npx skills add https://github.com/hai-series/hai-framework --skill hai-vecdb

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: hai-vecdb
Source: https://github.com/hai-series/hai-framework/tree/main/packages/cli/templates/skills/hai-vecdb
Command: npx skills add https://github.com/hai-series/hai-framework --skill hai-vecdb

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Unified interface to manage vector databases across LanceDB, pgvector, and Qdrant, enabling AI-powered retrieval workflows without vendor lock-in.

Core Features & Use Cases

Collection management (create, drop, exists, info, list) across backends
Vector CRUD and similarity search (insert, upsert, delete, search, count)
Use cases include RAG pipelines, semantic search, and knowledge retrieval integrated with embedding models

Quick Start

Initialize vecdb with your configuration and perform your first collection or vector operation.

Frequently Asked Questions about hai-vecdb

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

FAQPage Schema

How do I manage vector database operations across multiple backends without vendor lock-in?▼

Unified vector database operations across LanceDB, pgvector, and Qdrant prevent vendor lock-in by standardizing collection management and similarity search under a single interface for AI workflows.

What's the best way to perform similarity search for RAG pipelines across different vector databases?▼

Similarity search for RAG pipelines uses a unified interface to execute vector CRUD and retrieval across LanceDB, pgvector, and Qdrant, returning typed HaiResult APIs with consistent error handling.

Can I use a single API for collection management across LanceDB, pgvector, and Qdrant?▼

Yes, a single API handles collection management across LanceDB, pgvector, and Qdrant, allowing you to create, drop, check existence, retrieve info, and list collections uniformly.

Does this vector database interface support initialization and configuration validation for AI workflows?▼

The vector database interface supports initialization and configuration validation for AI workflows, ensuring backend setups are verified before performing vector CRUD and semantic retrieval operations.

How do I insert and upsert embeddings for semantic retrieval using a unified vector database interface?▼

Insert and upsert embeddings for semantic retrieval by calling the unified vector CRUD operations, which manage data consistently across LanceDB, pgvector, and Qdrant backends.

When should I use a unified vector database interface instead of a backend-specific client?▼

Use a unified vector database interface instead of a backend-specific client when building RAG pipelines or knowledge retrieval workflows that require the flexibility to switch between LanceDB, pgvector, and Qdrant without code changes.