Rate limits

Layered request limits, Retry-After and safe retry behaviour.

Rate limits are applied in Redis before the submission is processed. Malformed and spam attempts count toward the limits.

Payload limits

Request body
64 KiB
Fields per submission
50
Field name length
100
Scalar value length
10,000
List items
20
Idempotency key
128 chars

Rate limits

Per IP, per form
10/min
Per IP, all forms
60/min
Per form
120/min
Global
600/min
CORS preflight
120/min
Retention
90 days

429 response

{
  "success": false,
  "error": {
    "code": "rate_limited",
    "message": "Too many requests. Please try again later."
  },
  "request_id": "01JY8Z3K4M5N6P7Q8R9S0T1V2W"
}

The response includes Retry-After in seconds. Wait at least that long before retrying.

If Redis cannot safely apply the limit, the public API fails closed with 503 temporarily_unavailable rather than accepting unbounded traffic.

Retry guidance

  • Honour Retry-After.
  • Reuse the same idempotency key for the unchanged payload.
  • Use exponential backoff for 503.
  • Do not retry 403, 409 or 422 without fixing the request.
  • Do not reuse a Turnstile token.

See Handling rate limits and retries.