Stay inside every upstream quota

Rate Limits

Cap requests and tokens per minute or day on providers, models and individual provider keys, and let routing steer around them.

Your app → LLM Gateway
Your appLLM Gatewayopenai · key 1TPM cap hitopenai · key 262% of TPMazure · key 1standby
Capped keys leave rotation until the window slides
01

Requests & tokens

RPM, RPD, TPM and TPD windows.

02

Per provider key

Match each upstream account's quota.

03

Strict, soft or lax

Keep session pins past a cap if you choose.

04

Retry-After

Accurate reset hints on every 429.

Built for

  • Shared BYOK keys

    Spread traffic across several upstream keys

  • Tier-limited accounts

    Strict caps keep traffic inside token quotas

  • Long sessions

    Soft caps keep reasoning on the same key

More in routing & reliability

All features

Start using Rate Limits today

One API key. Every model. Live in minutes.