Skip to content

Rate limits

API v3 allows 40 points per key every 60 seconds: a read costs 1 point, a write 2.

Last updated

API v3 uses the same points budget as API v1, counted per API key.

Limit Value
Budget per window 40 points
Window length 60 seconds, fixed
Read (GET) 1 point
Write (POST, PATCH, DELETE) 2 points

A fresh window allows up to 40 reads or 20 writes. The MCP server has its own budget per store, with the same numbers; each tool call there costs 2 points.

Every response carries:

Header Meaning
X-RateLimit-Limit Points per window
X-RateLimit-Remaining Points left in this window
X-RateLimit-Reset When the window resets

The API answers 429 Too Many Requests with a Retry-After header and the same number of seconds in the body:

{
"errors": ["Rate limit exceeded: 40 points per 60 seconds (reads cost 1, writes cost 2). Retry after 17 second(s)."],
"code": "rate_limited",
"retry_after": 17
}

Wait that long, then retry. To stay under the limit, use list filters instead of reading records one by one, and use the bulk endpoints for many documents at once.

Was this page helpful?

Still stuck? Ask Quinn or chat with support