API Integration

Need to talk to Stripe, Twilio, OpenAI, or your own internal API? Mention it in the chat and you get a typed client back — with env vars, error handling, and a matching frontend shape — not just a raw fetch call.

What the generator sets up

• An integrations/ folder with one file per service.
• Environment variables in .env.example for keys and endpoints.
• Retry-with-backoff for transient errors, so a blip doesn't fail the request.
• A typed response shape on the frontend, regenerated when you change the model.

Example — Stripe checkout

"Add Stripe Checkout. On success, mark the order as paid and email the customer."

Result: a Checkout-session endpoint, a webhook handler that flips the order status, and an email template wired through your app's email provider.

Integrating your own API

1. Paste a sample response or an OpenAPI URL into the chat.
2. Say what you want — "generate a typed client for this and a hook to load orders on the dashboard."
3. You get the client, the types, and the UI wiring in one pass.

Handling keys and secrets

• The generator writes placeholders into .env.example; you fill the real values in .env.
• .env is git-ignored, so keys never reach the repo or the client bundle.
• Rotate any key you accidentally paste into code — treat committed secrets as public.

Common mistakes to avoid

• Pasting live secret keys into the chat. Describe the integration; add real keys yourself in .env.
• Calling a third-party API straight from the frontend. Route it through the backend so the key stays server-side.
• No error handling. Ask for retries and clear user-facing errors on network failures.

FAQ