tracert.site
The public face of Tracert — human routes, an agent entry point, capability pages and machine exports, all generated from the registry at build time. No hand-maintained marketing for a capability: every fact on a capability page comes from its TRACE Manifest.
Next.js (App Router), fully static. Zero UI dependencies; a small hand-rolled design system that adapts to light and dark.
Develop
npm ci
npm run dev # http://localhost:3000
npm run build # static export of every route
The site reads the registry from a sibling ../registry by default. Point elsewhere with TRACERT_REGISTRY_DIR. After changing manifests, rebuild — pages are baked at build time.
Routes
Human: / (routes fork for humans vs agents) · /about · /publish · /use · /capabilities · /capabilities/<id>
Machine:
| URL | Serves |
|---|---|
/llms.txt | This site, summarized for language models |
/index.json | The registry snapshot — every capability with status, pricing, media types, manifest URL |
/schemas/manifest/v0.1 | TRACE Manifest JSON Schema (served at its own $id) |
/schemas/receipt/v0.1 | TRACE Receipt JSON Schema (served at its own $id) |
/capabilities/<id>/manifest.json | One capability's canonical manifest, with an X-Tracert-Manifest-Hash header |
Deploy (Vercel)
The app lives in this site/ subdirectory of the monorepo. In the Vercel project, set Root Directory = site; the framework preset is Next.js. The build reads ../registry, which the registry-directory resolver locates from either the site root or the repository root — no extra configuration required. See ../DEPLOY.md.