---
updatedAt: 2026-03-13T00:44:32.000Z
agentTools:
  projectIndex: https://docs.split.io/llms.txt
---

# Rate Limiting

Our public API is guarded by a rate limiter using several techniques to prevent abuse. Calls to this API that are rate limited return a `429` status code along with relevant headers that indicate the remaining quota (`X-RateLimit-Remaining-Org` and `X-RateLimit-Remaining-IP`) and how long to wait until the next window specified in seconds (`X-RateLimit-Reset-Seconds-Org` and `X-RateLimit-Reset-Seconds-IP`). When rate limited, users should retry after a number of seconds equal to or greater than the maximum time of the last two `X-RateLimit-Reset-Seconds-XXX` headers.

These rate-limit headers are returned only by endpoints under `api.split.io`. Harness platform API endpoints (for example, endpoints under `api.harness.io`, such as `/user/getaggregatedusers`) do not return these headers. If you're migrating from deprecated Split API endpoints to Harness platform endpoints, you may notice that these headers are no longer included in responses.

> 🚧 API rate limiting and SDKs
>
> Rate limiting applies only to this public API and does not affect the API used by the SDKs to serve feature evaluations.

<br />