Rate limits
How many requests a workspace may send.
| Requests | Limit |
|---|---|
| All requests | 120 a minute per workspace |
Paid POSTs (/videos, /slideshows, /ai/*, retry) | 20 a minute per workspace |
The limits count every key of the workspace together, including connected apps like Claude: more keys do not raise them. An address that sends more than 10 wrong keys in a minute gets 429 rate_limited for a minute, even with a good key.
Every response says where you stand (RateLimit-Remaining is a best guess; the 429 itself is exact):
RateLimit-Limit: 120
RateLimit-Remaining: 117Over the limit, the API answers 429 rate_limited with Retry-After in seconds. Wait that long, then retry with the same Idempotency-Key.
Rate limits are technical and the same on every plan. The business limits (credits, the waiting queue, AI writes an hour, uploads a day) are in Credits & limits.
Polling politely
A video usually takes 1–3 minutes. Ask for a whole batch with GET /videos?batch_id=… in one call, about every 30 seconds plus 15 seconds for each video still waiting (never more often than that, at most 5 minutes apart).