Rate limits

Each API key has a per-minute request budget enforced server-side. The budget scales with plan:

Plan	Requests / minute
Free	30
Indie	120
Builder	500

When you’re over budget the server returns 429 RATE_LIMITED with a Retry-After header (seconds).

Headers

Every response — success or not — carries:

X-RateLimit-Limit:     <int>   // total budget
X-RateLimit-Remaining: <int>   // remaining this window
X-RateLimit-Reset:     <unix>  // seconds since epoch when the window resets

On a 429 we also include:

Retry-After: <seconds>

Recovering

The official retry strategy is exponential backoff with full jitter, capped at 60 seconds:

async function withBackoff<T>(fn: () => Promise<T>, attempts = 5): Promise<T> {
  let lastErr: unknown;
  for (let i = 0; i < attempts; i++) {
    try {
      return await fn();
    } catch (err: any) {
      lastErr = err;
      if (err?.code !== "RATE_LIMITED") throw err;
      const retryAfter = Number(err.retryAfter ?? 1);
      const cap = Math.min(retryAfter, 60);
      const wait = Math.random() * cap * 1000;
      await new Promise((r) => setTimeout(r, wait));
    }
  }
  throw lastErr;
}

Future versions of the official SDK will handle this automatically.