How to add SMS notifications to an AI-built app

AI coding tools — Claude Code, Cursor, ChatGPT — will happily scaffold a whole app for you. What they often leave out is the boring-but-important last mile: telling a real person that something happened. SMS is the simplest way to close that gap, and because Bremlio is a plain HTTP endpoint, it drops into whatever stack the AI generated.

Why SMS is the easy add-on

No SDK, no build step, no platform account for your users. Any language that can make an HTTP POST can send a text. That means you can paste one function into a Next.js route, a Python worker, a Supabase edge function, or a cron job — and it just works.

One function, any stack

// Works anywhere fetch exists (Node 18+, edge runtimes, Deno)
export async function notify(to, message) {
  const r = await fetch("https://api.bremlio.com/v1/messages", {
    method: "POST",
    headers: {
      "X-Api-Key": process.env.BREMLIO_API_KEY,
      "Content-Type": "application/json"
    },
    body: JSON.stringify({ to, message })
  });
  return r.json();
}

Good prompts for your AI tool

If you're pairing with an AI, be specific. Prompts like these get clean results:

• "Add a server-side function that POSTs to https://api.bremlio.com/v1/messages with an X-Api-Key header and a JSON body of to and message."
• "Read the API key from an environment variable, never hardcode it."
• "Call it when an order status changes to 'shipped'."

Two things the AI usually forgets

• Keep the key on the server. Never call the SMS API from the browser — the key would be exposed.
• Handle the non-200 case. Check the response so a failed notification doesn't silently disappear.