Per-attempt limits that suit agents

Request Timeouts

Per-attempt time limits that reset on every retry and never count your tool execution time.

Get startedRead the docs
Agent call → Per-attempt timer
Agent callPer-attempt timerAttempt 1limit reachedAttempt 2fresh windowTool runnot counted
Every retry starts a new window
01

Per attempt

Retries and fallbacks get fresh time.

02

Tools excluded

Your tool time never counts.

03

Agent friendly

Hours-long runs are unaffected.

04

Tunable

Lower per project, or set when self-hosting.

Built for

  • Agent loops

    Many short calls over long sessions

  • Strict SLAs

    Cap slow attempts per project

  • Self-hosting

    Set limits with env variables

More in routing & reliability

All features

Service Tiers

Trade latency for cost, per request

Caching

Pay once for repeated work

Custom Providers

Bring any OpenAI-compatible endpoint

Start using Request Timeouts today

One API key. Every model. Live in minutes.

Start buildingBrowse models