Show HN: Quoth – Semantic search for quotes using pgvector and OpenAI embeddings
quoth.appI built Quoth (https://quoth.app) to solve a personal problem: finding quotes by meaning rather than exact words. You can search for "adventure" and discover quotes about exploration you've never seen before.
The project uses Next.js 15 with App Router, TypeScript, Prisma 6, PostgreSQL with pgvector for semantic search, OpenAI embeddings for quote vectors, Dynamic OG image generation for sharing. It was built in ~3 weeks using AI-assisted development (Claude/Copilot).
Features include semantic search across ~1000 curated quotes, community submissions with verification system, daily newsletter, social sharing with custom typography.
This is a passion project with no business model. I'm particularly interested in feedback on the search relevance and any ideas for improving the curation process.
Source isn't public yet but considering it. Happy to discuss the AI-assisted development workflow that made this feasible as a solo side project.
No comments yet.