First Pattern to Production

PreviousNext

A practical path from choosing a pattern to shipping it safely in production.

This is the recommended rollout sequence if you are new to the platform.

Step 1: Choose import path

  • Use Install with CLI if you already have an app.
  • Use Download as App if you want full scaffolding and setup defaults.

Step 2: Get implementation context fast

Use Copy skill files from the pattern viewer and paste into your AI coding tool.

Ask the AI for focused changes:

  • adapt model/provider
  • customize tools and schemas
  • keep external API surface unchanged

Step 3: Configure runtime

Set required environment variables and external service credentials before testing.

Step 4: Validate locally

pnpm lint
pnpm typecheck
pnpm dev

Verify:

  • route loads
  • streaming/tool calls succeed
  • error states render correctly

Step 5: Production hardening

  • add rate limits where applicable
  • add monitoring/logging
  • verify auth boundaries
  • test failure fallbacks for external APIs

Step 6: Ship incrementally

  • deploy behind internal feature flag
  • gather traces and error rates
  • then roll out broadly