Rate limits

240 requests per 15 minutes, counted per key.

How the allowance is counted

Per key, not per business and not per IP address.

Because the count is per key, one busy integration cannot exhaust another's allowance, and several of your servers sharing one outbound IP address are not charged for each other's traffic. If one system genuinely needs more headroom than another, give it its own key.

Response headers

RateLimit-Policy: 240;w=900
RateLimit-Limit: 240
RateLimit-Remaining: 237
RateLimit-Reset: 842

RateLimit-Reset is the seconds remaining in the current window. Once the allowance is spent, requests answer 429 with a Retry-After header carrying the same number of seconds — wait that long rather than retrying immediately.

Sizing a catalogue sync

A full catalogue push costs one request per item. At 240 per 15 minutes, a catalogue of a few hundred items fits comfortably in one window; a few thousand should be spread across several, or reduced to a nightly full push plus PATCH calls for the changes in between.