Rate limits
What the limits are, how they are communicated, and how to stay inside them.
Limits exist because searches cost suppliers money too, and a look-to-book ratio that runs away will get your supplier contracts reviewed. The limits below are the sandbox defaults; production limits are set per account against your forecast.
| Surface | Sandbox default | Window |
|---|---|---|
| Token endpoint | 30 requests | per minute |
| Search endpoints | 60 requests | per minute |
| Content and rate checks | 300 requests | per minute |
| Booking writes | 30 requests | per minute |
| Read endpoints | 600 requests | per minute |
Response headers
Headers
X-RateLimit-Limit: 60X-RateLimit-Remaining: 41X-RateLimit-Reset: 1756377660Retry-After: 19- On 429, honour
Retry-After. Retrying sooner extends the window rather than shortening it. - Cache tokens. Token-endpoint pressure is the most common cause of a 429 in a new integration.
- Cache static content (property descriptions, images, amenities). It changes rarely and does not need to be fetched per search.
- Do not poll booking status. Use webhooks.