Documentation menu

api

Rate limits

Current fair-use posture and the published limits that arrive with GA infrastructure.

Last updated 2026-07-10 · beta

Today

The beta runs on fair use without hard per-key limits at the application layer. Concurrency is proven to 12 simultaneous multi-shot sequences in the load gate; sustained abusive traffic may be throttled at the platform layer.

What arrives with GA

  • Published per-key request and concurrency limits by tier.
  • 429 with a Retry-After header when exceeded.
  • Higher limits on the developer and agency tiers.

When limits ship they will be documented here and announced in the changelog.

Rate limits | Cadenza Docs