I started type.lol in 2015 with a friend as a simple list of independent type foundries — basically a styled airtable doc. I'm a designer and I kept running into the same problem: I'd want to explore type beyond the usual distributors, end up with 30 tabs open, and lose track of what I'd already looked at. The list helped but it wasn't enough.

I've since rebuilt it from scratch as a proper app. It now indexes 800+ foundries from 61 countries, 14k typefaces, and 1,500 designers. No signup required to browse. No ads, no sales, no commissions — it embeds actual foundry websites in iframes so traffic goes directly to them.

You can browse by carousel, list, virtualized table, force-directed graph (showing relationships between foundries/designers/typefaces), or a 3D globe plotting foundries by location. Filter by classification, variable fonts, trial availability, country, designer, language support, or release era.

Some technical details since this is HN:

- React 19 + Vite + TypeScript + Zustand + Supabase - Three-tier data loading: IndexedDB cache (instant) → Supabase paginated fetch → static JSON fallback. Background staleness check. - @tanstack/react-virtual for carousel and table views across 14k typefaces - Client-side ranked search with related-entity expansion (match a designer → surface their foundries and typefaces) - Graph view: react-force-graph-2d. Globe: react-globe.gl + three.js - 577 of 779 foundry sites allow iframe embedding. Rest fall back to OG images or screenshots.

On the data pipeline: I scrape foundry websites, then use Claude Haiku to extract structured metadata (classifications, designer credits, language support, weights, etc.) from the page content. It's not perfect — some classifications are wrong, some designer credits are incomplete — but it gets me to ~80% and I manually correct from there. The enriched data goes through a build script into static JSON and syncs to Supabase.