MemoryCV

AI-powered resume builder from your memory export. TanStack Start + React 19, bilingual (EN/KU), vector PDF export.

Setup

npm install
cp .env.example .env.local
# Set GEMINI_API_KEY in .env.local
npm run dev

Scripts

Command	Description
npm run dev	Development server
npm run typecheck	TypeScript
npm run test	Vitest unit tests
npm run lint	ESLint
npm run build	Production build
npm run check	typecheck + test + build

Deploy (Vercel)

1. Connect repo to Vercel
2. Set GEMINI_API_KEY in project environment variables
3. Auth & cloud: VITE_CLERK_PUBLISHABLE_KEY, CLERK_SECRET_KEY, VITE_SUPABASE_URL, VITE_SUPABASE_ANON_KEY — see docs/SUPABASE_CLERK_SETUP.md
4. Email (optional): RESEND_API_KEY, RESEND_FROM_EMAIL for welcome emails
5. Do not set ALLOW_CLIENT_API_KEY in production

See docs/SECURITY.md, CONTRIBUTING.md, and SMOKE_CHECKS.md.

Deploy target: Vercel (Nitro preset). Cloudflare wrangler.jsonc is archived under docs/archive/.

Architecture

• Client state: Zustand + localStorage (memorycv-store; API keys are not persisted)
• AI: Server functions in src/lib/ai.functions.ts → Gemini API
• Export: Vector PDF via dom-to-svg + svg2pdf; raster PDF via html-to-image
• Optional cloud: Supabase auth + user_data table