Color Finder

A semantic color search engine powered by vector embeddings. Search for colors using natural language descriptions and get matching color suggestions.

Features

• Natural language color search using OpenAI embeddings
• 30,000+ named colors in the database
• Fast vector similarity search using PostgreSQL/Supabase
• Returns color names, hex codes, and similarity scores

Technical Decisions (All Lightly Taken)

• Uses OpenAI's text-embedding-3-small model for semantic embeddings
• PostgreSQL vector extension with IVFFlat indexing
• Supabase for database hosting and RPC functions
• TypeScript/Node.js backend

Getting Started

1. Clone the repo
2. Clone the .env.example to .env and swap in your OpenAI / Supabase credentials
3. Install dependencies: npm install
4. Run the development server: npm run dev

Credits

Written by lui for brandmint.ai Color data sourced from meodai/color-names