Back
AI Holiday Recommendation App

AI Holiday Recommendation App

#RAG #OpenAI-embeddings #Supabase #vector-database

Island Explorer is a proof-of-concept application demonstrating Retrieval-Augmented Generation (RAG) in action. The app transforms custom island destination data into OpenAI embeddings, stores the chunked vectors in Supabase's PostgreSQL database with vector extensions, and enables users to discover their perfect island getaway through natural language queries.

The recommendation workflow is seamless: users submit their travel preferences via a simple form, the backend converts these into vector embeddings, Supabase performs a similarity search using RPC functions to find matching island data chunks, and GPT-4 synthesizes the results into personalized, conversational recommendations — all powered by a lightweight serverless architecture on Netlify.

Visit Live Website
NexaHub Project

Key Features

RAG Pipeline

Implements a complete Retrieval-Augmented Generation workflow — user preferences are embedded, matched against stored vectors, and enriched by GPT-4 to produce grounded, context-aware recommendations.

Supabase Vector Storage

Leverages Supabase's PostgreSQL with vector extensions to store chunked island data as embeddings, enabling fast semantic similarity searches via custom RPC functions.

Serverless Architecture

Backend logic runs on Netlify Functions, providing a scalable, zero-infrastructure API layer that handles embedding generation, vector queries, and AI response synthesis.

Lightweight Frontend

Built with vanilla JavaScript, CSS, and HTML — a clean, dependency-free interface that collects user preferences and displays AI-generated island suggestions.

Visit Live Website

Ready to Start Your Project?

Get Started Back