Music search. Finally.
Every AI asked about music guesses — confidently, fluently, wrongly. There's no structured, agent-ready music data layer on the internet. So we're building one.
Dig is a search engine and data layer built on the full Discogs CC0 dataset. 24 million records. 2.5 million masters. 580,000 artists. Every entity cross-linked — artists to releases, releases to credits, credits to labels. Click anything, follow the thread.
- Deep catalog search — full-text search across 24M+ records, fast on mobile, dark mode
- Connected entity pages — every artist, label, release, and master cross-linked with type badges, filter chips, and newest-first sorting
- Enrichment layers — MusicBrainz crosswalks, Wikidata context (bios, locations, genres), setlist.fm performance history, Cover Art Archive artwork
- No AI in the data path — deterministic retrieval only, structured data, no hallucinations
TypeScript · Postgres · Next.js · Fastify · Kysely · Redis · Fly.io
Early stage. Active alpha. Building in public.
See the build log and about page for more.
Built by @b1rdmania.