# Tracker Dashboard — Vercel Deployment Roadmap

> **App:** `@bytelyst/tracker-web`
> **Repo:** `learning_ai_common_plat` → `dashboards/tracker-web/`
> **Status:** ✅ Ready

---

## Current State

- **Framework:** Next.js 16.1.6, React 19.2.3, TailwindCSS v4
- **Output:** Vercel-aware (`process.env.VERCEL ? {} : { output: 'standalone' }`)
- **Dependencies:** 6 `@bytelyst/*` packages via `workspace:*` (monorepo)
- **Build:** `next build --webpack`
- **Server-side:** AKV secret resolution only (via instrumentation.ts)
- **No `file:` references**

## Deployment Steps

### Step 1: Create Vercel Project (~5 min)

- [ ] Connect `learning_ai_common_plat` repo to Vercel (or reuse org if admin-web already connected)
- [ ] Set **Root Directory** to `dashboards/tracker-web`
- [ ] Framework preset: **Next.js**

### Step 2: Configure Environment Variables (~5 min)

- [ ] `JWT_SECRET` — JWT signing secret
- [ ] `AZURE_KEYVAULT_URL` — Azure Key Vault URL (optional)
- [ ] `PLATFORM_SERVICE_URL` — URL of platform-service API
- [ ] `NODE_ENV` = `production`

### Step 3: Verify Build (~5 min)

- [ ] Trigger deploy
- [ ] Verify public roadmap page loads (`/roadmap`)
- [ ] Test item submission and voting
- [ ] Verify security headers

### Step 4: Domain Configuration (~5 min)

- [ ] Add custom domain (e.g., `tracker.bytelyst.com`)

## Notes

- **pretest script:** Has a `pretest` script that builds workspace deps — Vercel won't run this, but `pnpm build` in the root handles workspace resolution.
- **Lighter footprint:** Fewer server-side deps than admin-web. Mostly a client-side app that calls platform-service APIs.
- **Bundle size:** Has `bundlesize` checks configured — good for monitoring after deploy.

## Estimated Total Effort: ~20 minutes