Recipes
End-to-end walkthroughs for the scenarios people actually face.
Each recipe is self-contained: what to run, in what order, what to verify between steps. They assume the quickstart is done (CLI installed + signed in), and link to relevant guides for deeper context.
Ship something now
MVP demo
Investor wants a link, you have five minutes.
Client delivery
Demo's tomorrow, deploy tonight. Per-client URL, time-bound.
Hackathon demo
Build it, deploy it, QR it for the audience.
Routine workflows
CI/CD with GitHub Actions
Auto-deploy on every push to main, smoke check, fail loud.
Per-PR previews
Channel per PR, comment the URL, auto-clean on close.
Daily reports
Cron-driven HTML reports, same URL, always current.
Monorepo deploys
One repo, many apps, per-package CI matrix.
Migrations
From Vercel
Move your app and its domain. Hour-of-work, not day-of-work.
From Netlify
Same shape, different platform. _redirects + _headers.
From Replit
$700/mo → $0/mo for hosting a static export.
Patterns
Deploy from AI canvas
One click inside ChatGPT, Claude, Gemini, v0, or AI Studio.
Full agent workflow
Claude / Cursor builds + deploys + shares — end to end.
Share with clients
Private app + time-bound share link, no sign-up.
Don't see your scenario?
The CLI accepts --json, the API is fully documented, and the MCP server exposes the same surface as tools. You can compose anything these recipes do.
- CLI reference — every command
- API reference — every endpoint
- GitHub Discussions — request a recipe