Does this skill require a specific vector database?

No, it uses the pgvector extension for PostgreSQL, allowing you to store vectors directly in your existing relational database without needing a separate vector-only provider.

What embedding models are supported?

By default, it is configured to use Google's text-embedding-004 via AI Gateway, but it can be easily customized to use any embedding model supported by the Vercel AI SDK.

How does the chunking mechanism handle long documents?

The skill uses a recursive text splitter that respects paragraph and sentence boundaries. It breaks text into ~512 token segments with a 200-character overlap to ensure context is maintained across chunks.

What are the requirements for installation?

This skill requires a Next.js App Router project with the 'db' (Drizzle), 'ai-core', and 'ai-rag-ingest' skills already installed and configured.

AI RAG Vectors & Semantic Search

Name: AI RAG Vectors & Semantic Search
Author: mattwoodco

bymattwoodco

•

데이터 과학 및 ML

Integrates a vector embedding layer into Next.js applications using pgvector and AI Gateway for high-performance semantic search and retrieval.

The AI RAG Vectors skill provides a robust infrastructure for building Retrieval-Augmented Generation (RAG) systems within Next.js environments. It automates the complex process of recursive text chunking with configurable overlap, generates high-quality embeddings via AI Gateway, and stores them in a Postgres database powered by pgvector. With built-in support for cosine similarity search and HNSW indexing, it enables developers to transform static document pages into a searchable knowledge base, allowing AI agents to retrieve relevant context with high precision.

주요 기능

013 GitHub stars

02Seamless pgvector integration for Postgres-native vector storage

03Automatic HNSW indexing for optimized retrieval at scale

04Recursive text chunking with context-preserving overlap

05High-performance semantic search using cosine similarity

06Batch embedding generation via AI Gateway and Vercel AI SDK

사용 사례

01Building a document-based Q&A system or internal knowledge base

02Providing context-aware data retrieval for AI chatbots and agents

03Enabling semantic search across large PDF or text datasets

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add mattwoodco/skills ai-rag-vectors

For use in Claude.ai and ChatGPT

Download Skill