for machines

AI agents

This site is a static mirror of the sh4dcn directory. Every listing has a structured twin — read those instead of scraping the UI. Pro/paid tiers are indexed as metadata (URLs) only.

Synopsis

# sh4dcn-v2 — instructions for AI coding agents
# static mirror of sh4dcn · 253 registries · generated 2026-07-28

1. Read the directory:   curl -s https://sh4dcn-v2.vercel.app/json
2. Read a library brief: curl -s https://sh4dcn-v2.vercel.app/md/7ovr.md
3. Install a registry:   npx shadcn@latest add <registry_json_url>
4. Never scrape the HTML — every page here has a machine-readable twin.

Endpoints

/json json Full merged directory — every library + component with URLs, install commands, tags, pro info, and facet indexes.
/md/index.md markdown Index of one agent-readable brief per library (frontmatter metadata, install commands, full component table).
/md/7ovr.md markdown Example per-library brief (@7ovr).
/csv/libraries.csv csv One row per library: slug, category, tags, description, URLs, pro info, artifact paths.
/csv/components.csv csv One row per component across all libraries.
/csv/files.csv csv Index of every published machine-readable artifact.
/llms.txt text AI-agent entry point (canonical data host: sh4dcn.vercel.app).
/sitemap-index.xml xml Sitemap of every prerendered page.

Canonical source

Data originates from sh4dcn.vercel.app (refreshed daily from the official shadcn/ui v4 registry directory) and is mirrored here at build time. For always-freshest data, prefer the v1 host; for zero-JS static reads, this mirror is equivalent.