Since ~2026-07-16, every deployment of our Next.js 16.2.10 frontend returns Vercel's platform 404 (x-vercel-error: NOT_FOUND) on every path of every hostname, before requests reach the app. A Python serverless project on the same team serves normally throughout.
Team: jan1423 (team_IVzEOWKRLIERtWNg2ZkpCGkd)
Affected: nclex-ai-platform-web (prj_KLRWFW1u05IRSEMyHoCVH9OdDWXm) and emberprep (prj_cRgCW0sVzSiXNCbkSw4WG43Swubt, created fresh today from the same repo — identical failure on its first-ever deploy)
Unaffected control: nclex-ai-platform (Python, same team, serves 200s)
Evidence:
- Immutable deployments changed behavior:
dpl_4KEPTVwUQ8YrtSLUNL5YGCPNvR1Mwas browser-verified working on Jul 15; today its own deployment URL 404s on every path with no redeploy in between. - Current prod
dpl_235pTHaaozvgDbQx8Y5DT1TzDTAv: build log clean, all routes listed, READY/PROMOTED, aliasError null — yet static assets (/icon.svg) and public routes all 404. - Reproduced from fra1 (client) and iad1 (server-side fetch) — not regional.
- Tried without success: no-cache rebuild, domain remove/re-add, explicit re-alias, promote (409 already-current), disabling Deployment Protection, migrating middleware.ts to proxy.ts, and a brand-new project.
- Control plane healthy via API (READY/PROMOTED, aliasAssigned, not paused); zero function invocations logged — requests never reach deployments.
Symptoms match threads 45317 and 45939 (also multiple projects incl. a fresh clean project on a Hobby team). Looks like team-level edge-routing state only Vercel can repair. Please restore serving for team jan1423 or advise.
