Errors & rate limits

Error shape

Many endpoints return JSON with a human-readable message and an optional errCode:

{
  "message": "Unauthorized",
  "errCode": 401
}

Some endpoints also include an error string with more detail.

Common status codes

Status	Meaning
200 / 201	Success
202	Accepted — job queued (poll with the returned jobId)
400	Invalid request (missing fields, bad URL, prompt limits, etc.)
401	Missing or invalid API key
403	Plan or credit limit — upgrade or wait for credits
404	Resource or job not found
429	Too many requests from this IP
500	Server error

Rate limits

API traffic is limited per client IP:

Limit	Window
1500 requests	15 minutes

After about 1000 requests in the same window, responses may be delayed (slowdown) before hard limiting applies.

If you hit the cap, wait for the window to reset and retry with backoff.

Credits and plan limits

Generation endpoints consume AI credits according to your plan. When you are out of credits you typically receive 403 with a plan/limit message. Visibility report endpoints also enforce per-plan prompt counts.