About this collection

Explore is a public, searchable index of a curated ML research collection: papers, technical blogs, models, code repos, and datasets. Every entry is individually researched — what it is, what it claims, how to access it, and under what license — before it lands here.

How entries get added

The collection is fed by two Google Sheets that act as the database. New items start as shared links (papers, repos, datasets spotted on X, in Discords, or elsewhere), get researched and verified, and are appended as structured rows. This site pulls those sheets and rebuilds itself daily, so anything added to a sheet shows up here within 24 hours — no manual publishing step.

Searching

• Full-text search across titles, authors, summaries, notes, keywords, and links.
• Filter papers by type (paper, blog, code, model, thread) and by year.
• Sort by newest or by title. Every entry has a permanent URL you can share.

Data & freshness

The site is a read-only view: the sheets remain the source of truth. Each build records its sync time, shown on the front page. If a link rots or a fact is wrong, the fix happens in the sheet and flows through on the next sync.

Stack

Built with Next.js on Vercel. Sheet data is fetched at build time and baked into static pages; search runs entirely in your browser with MiniSearch, so queries are instant and private — nothing you type leaves your device.