Job Command Center

An AI-powered job application tracker and PWA — built to eliminate the spreadsheet and put your entire job search in one place.

---

What it does

• Upload once — AI parses your resume into skills, experience, and a strength score
• Daily job discovery — aggregates listings from Remotive, Arbeitnow, and HN Who's Hiring; scores each one 0–100 against your resume
• Kanban pipeline — drag-and-drop board tracks every application from Saved → Offer / Rejected
• AI cover letters — streams a tailored letter word-by-word; three tone options; version history
• Interview prep — generates likely questions by category with hints, key themes, and gap analysis
• Push notifications — daily digest at 9am, follow-up reminders you set per application
• Works offline — installable PWA with service worker caching

---

Tech Stack

Layer	Technology
Framework	Next.js 14 App Router + TypeScript
Styling	Tailwind CSS + shadcn/ui (dark-first)
Auth	Clerk v6 (Google OAuth + magic link)
Database	PostgreSQL via Supabase + Prisma 7
AI	Vercel AI SDK + Google Gemini 1.5 Pro/Flash
File Storage	Cloudflare R2 (S3-compatible)
State	TanStack Query v5 (server state)
Drag-and-drop	dnd-kit
Charts	Recharts
Animations	Framer Motion
Push	web-push (VAPID)
PWA	next-pwa (Workbox)
Hosting	Vercel (cron jobs + edge)

---

Features

• Discover — scan 3 free job sources, filter by score / remote, paginated with sort
• AI Scoring — Gemini Flash scores each job vs your active resume, flags strengths & gaps
• Pipeline Kanban — 6 default stages, drag-and-drop, click-to-open detail panel
• Cover Letter Generator — streaming SSE output, tone selector, 5-version history
• Interview Prep — AI questions by category (Behavioural / Technical / Situational / Culture)
• Follow-up Reminders — set a date on any application, get a push notification
• Analytics — response rate, funnel chart, weekly trend, average score
• Dark / Light / System theme toggle
• PWA — installable on iOS and Android, offline read access

---

Local Setup

# 1. Clone and install
git clone <repo>
cd job-command-center
npm install

# 2. Copy env template
cp .env.example .env.local
# Fill in all values (see table below)

# 3. Push the DB schema
npx prisma db push

# 4. Run
npm run dev

Open http://localhost:3000.

---

Environment Variables

Variable	Where to get it
DATABASE_URL	Supabase → Settings → Database → Connection string (Transaction mode)
DIRECT_URL	Supabase → Settings → Database → Connection string (Session mode)
NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY	clerk.com → API Keys
CLERK_SECRET_KEY	Clerk → API Keys
GOOGLE_GENERATIVE_AI_KEY	aistudio.google.com → Get API key
CLOUDFLARE_R2_ACCOUNT_ID	Cloudflare → R2 → Overview
CLOUDFLARE_R2_ACCESS_KEY_ID	Cloudflare → R2 → Manage API tokens
CLOUDFLARE_R2_SECRET_ACCESS_KEY	Same as above
CLOUDFLARE_R2_BUCKET_NAME	Your R2 bucket name
CLOUDFLARE_R2_PUBLIC_URL	Cloudflare → R2 → Bucket → Settings → Public Development URL
UPSTASH_REDIS_REST_URL	upstash.com → Redis → REST API
UPSTASH_REDIS_REST_TOKEN	Same as above
VAPID_PUBLIC_KEY	Generate: npx web-push generate-vapid-keys
VAPID_PRIVATE_KEY	Same command
NEXT_PUBLIC_VAPID_PUBLIC_KEY	Same as VAPID_PUBLIC_KEY
NEXT_PUBLIC_APP_URL	http://localhost:3000 locally, your Vercel URL in prod
CRON_SECRET	Any random string — set same value in Vercel env

---

Architecture

• API Routes — all server logic lives in app/api/. Auth via Clerk middleware (middleware.ts)
• AI layer — lib/ai.ts exports proModel (Gemini 1.5 Pro) and flashModel (Flash). Swap to Claude by changing two lines
• DB — Prisma 7 with pg adapter. Schema in prisma/schema.prisma, datasource URL in prisma.config.ts
• File uploads — resumes go to Cloudflare R2, URL stored in DB
• Crons — vercel.json schedules daily digest (09:00 UTC) and reminder check (08:00 UTC)
• Push — VAPID keys + web-push. Subscriptions stored in PushSubscription table

---

What this demonstrates

• Full-stack ownership — DB schema → serverless API → animated PWA UI, shipped solo
• AI product thinking — streaming UX, structured output, prompt caching strategy, model swap path
• PWA expertise — service workers, background sync, web push, offline-first
• Realtime patterns — SSE streaming for cover letters, optimistic Kanban updates
• Infrastructure — Vercel cron jobs, R2 object storage, Supabase RLS, edge deployment