wrangler.jsonc

{
  "$schema": "./node_modules/wrangler/config-schema.json",
  "name": "hacksv-dashboard",
  "account_id": "3d3da26ef3b03d928bc5c29805655f9a",
  "compatibility_date": "2026-07-09",
  "compatibility_flags": ["nodejs_compat"],
  "main": "./worker/index.ts",
  // Reserve these two path prefixes on the hack.sv zone. Everything the app
  // needs (SPA pages, built assets, and the API) lives under them.
  //
  // Trailing-wildcard patterns are REQUIRED: Cloudflare route matching includes
  // the query string, and `?` is not allowed in patterns — so "hack.sv/auth"
  // would NOT match "/auth?email=…" (it would fall through to the apex site).
  // Note this also captures /auth* and /dashboard* prefixes (e.g. /authors),
  // so the apex site must not use those prefixes for its own pages.
  "routes": [
    { "pattern": "hack.sv/auth*", "zone_name": "hack.sv" },
    { "pattern": "hack.sv/dashboard*", "zone_name": "hack.sv" }
  ],
  "assets": {
    // Serve the built React SPA. Anything that isn't a real asset falls back
    // to index.html (client-side routing), EXCEPT the auth endpoints below,
    // which always hit the Worker first. The /auth login page and everything
    // else fall through to the SPA. Assets are based under /dashboard/.
    "not_found_handling": "single-page-application",
    "run_worker_first": [
      "/auth/session",
      "/auth/google",
      "/auth/callback",
      "/auth/magic/*",
      "/auth/verify-email",
      "/auth/verify-resend",
      "/auth/logout",
      "/auth/profile",
      "/auth/registration",
      "/auth/registration/*",
      "/auth/connect/*"
    ]
  },
  "d1_databases": [
    {
      "binding": "DB",
      "database_name": "hack-sv-dashboard",
      "database_id": "fbb87c66-6036-42d3-9406-a27c9942f99f",
      // Use the real remote D1 in `vite dev` too, instead of an empty local
      // database.
      "experimental_remote": true
    }
  ]
}