Resume Console

An AI resume consultant that writes, formats, and polishes your resume in minutes. Just talk — it handles the rest.

Features

• AI-guided writing — Chat with an AI consultant that asks the right questions and writes compelling resume content
• Live preview — Watch your resume update in real-time as you chat
• PDF upload — Upload an existing resume PDF and the AI extracts your details
• PDF generation — One-click PDF rendering via Cloudflare Browser Rendering
• Multiple templates — Clean, Professional, and Bold styles
• Dark mode — Full dark theme support
• Credit-based pricing — Stripe integration, no subscriptions

Stack

• Frontend: Astro + React 19 + TanStack Router/Query
• Auth: BetterAuth with D1 adapter
• AI: pi-ai with MiniMax M2.5 via OpenRouter
• Database: Cloudflare D1 (SQLite)
• Storage: Cloudflare R2
• PDF: Cloudflare Queues + Browser Rendering API
• Payments: Stripe Checkout
• Email: Resend
• Hosting: Cloudflare Pages

Setup

Prerequisites

• Node.js 22+
• pnpm
• A Cloudflare account
• Wrangler CLI: npm install -g wrangler

1. Clone and install

git clone https://github.com/tinyfatco/resume-console.git
cd resume-console
pnpm install

2. Create Cloudflare resources

# D1 database
wrangler d1 create resonate-db
# Copy the database_id into wrangler.toml and workers/resonate/wrangler.toml

# R2 bucket
wrangler r2 bucket create resonate-storage

# Queue
wrangler queues create pdf-generation
wrangler queues create pdf-generation-dlq

3. Run migrations

# Local
wrangler d1 migrations apply resonate-db

# Production
wrangler d1 migrations apply resonate-db --remote

4. Configure environment

cp .env.example .dev.vars
# Fill in your API keys

For production, set secrets via:

wrangler pages secret put BETTER_AUTH_SECRET
wrangler pages secret put OPENROUTER_API_KEY
wrangler pages secret put RESEND_API_KEY
wrangler pages secret put STRIPE_SECRET_KEY
wrangler pages secret put STRIPE_WEBHOOK_SECRET
wrangler pages secret put STRIPE_PUBLISHABLE_KEY

5. Run locally

pnpm dev

6. Deploy

# Main app (CF Pages)
pnpm build && wrangler pages deploy dist --project-name=resume-console

# PDF worker (CF Workers)
cd workers/resonate && wrangler deploy

Architecture

src/
├── components/app/     # React SPA (TanStack Router)
│   ├── ChatPane.tsx    # Chat interface with AI
│   ├── PreviewPane.tsx # Live resume preview (iframe)
│   └── ResumeSidebar.tsx
├── lib/
│   ├── ai.ts           # pi-ai model config
│   ├── auth.ts         # BetterAuth setup
│   ├── stripe.ts       # Stripe + credit tiers
│   └── tools.ts        # AI tool definitions (create/edit/list resume, bash)
├── pages/api/          # Astro API routes
│   ├── chat.ts         # SSE streaming endpoint
│   ├── resumes/        # CRUD
│   ├── credits/        # Stripe checkout + balance
│   └── auth/           # BetterAuth handler
├── middleware.ts        # Auth middleware
└── prompts/
    └── resume-builder.md  # System prompt

workers/resonate/       # PDF generation worker
    └── src/index.ts    # Queue consumer → Browser Rendering → R2

templates/              # HTML resume templates
    ├── clean.html
    ├── professional.html
    └── bold.html

How It Works

1. User sends a message via the chat pane
2. The message is streamed to OpenRouter (MiniMax M2.5) with the resume consultant system prompt
3. The AI can call tools: create_resume, edit_resume, list_resumes, read_resume, archive_resume, and bash
4. Tool results update the resume in D1, and the preview pane refreshes via polling
5. "Generate PDF" enqueues a job → the PDF worker renders HTML with headless Chrome → stores in R2

Tests

pnpm test

License

MIT